EXCLUDE_FROM_ALL means something else for add_lit_testsuite as it does
for something like add_executable. Distinguish between the two by
renaming the variable and making it an argument to add_lit_testsuite.
Differential revision: https://reviews.llvm.org/D74168
Fixes [[ https://bugs.llvm.org/show_bug.cgi?id=44733 | TEST 'libomp :: ompt/synchronization/reduction/tree_reduce.c' FAILED on 32-bit x86 ]]
For 32-bit we need at least 3 variables to avoid atomic reduction to be
choosen by runtime function `__kmp_determine_reduction_method`.
This patch adds reduction variables to the testcase.
Reviewers: mgorny, Hahnfeld
Differential Revision: https://reviews.llvm.org/D73850
Summary:
[nfc][libomptarget] Remove SHARED annotation from local variables
A few local variables in reduction.cu were marked SHARED. This patch leaves
all per-kernel global state localised in omp_data.cu.
Reviewers: ABataev, jdoerfert, grokos
Reviewed By: jdoerfert
Subscribers: openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D73239
Summary:
This patch is to fix issue in the following simple case:
#include <omp.h>
#include <stdio.h>
int main(int argc, char *argv[]) {
int num = omp_get_num_devices();
printf("%d\n", num);
return 0;
}
Currently it returns 0 even devices exist. Since this file doesn't contain any
target region, the host entry is empty so further actions like initialization
will not be proceeded, leading to wrong device number returned by runtime
function call.
Reviewers: jdoerfert, ABataev, protze.joachim
Reviewed By: ABataev
Subscribers: protze.joachim
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D72576
Summary:
[libomptarget] Implement smid for amdgcn
Implementation is in a new file as it uses an intrinsic with
complicated encoding that warranted substantial comments.
Reviewers: jdoerfert, grokos, ABataev, ronlieb
Reviewed By: jdoerfert
Subscribers: jvesely, mgorny, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D72956
The reference counter for global objects marked with declare target is INF. This patch prevents the runtime from incrementing /decrementing INF refcounts. Without it, the map(delete: global_object) directive actually deallocates the global on the device. With this patch, such a directive becomes a no-op.
Differential Revision: https://reviews.llvm.org/D72525
TSan spuriously reports for any OpenMP application a race on the initialization
of a runtime internal mutex:
```
Atomic read of size 1 at 0x7b6800005940 by thread T4:
#0 pthread_mutex_lock <null> (a.out+0x43f39e)
#1 __kmp_resume_64 <null> (libomp.so.5+0x84db4)
Previous write of size 1 at 0x7b6800005940 by thread T7:
#0 pthread_mutex_init <null> (a.out+0x424793)
#1 __kmp_suspend_initialize_thread <null> (libomp.so.5+0x8422e)
```
According to @AndreyChurbanov this is a false positive report, as the control
flow of the runtime guarantees the ordering of the mutex initialization and
the lock:
https://software.intel.com/en-us/forums/intel-open-source-openmp-runtime-library/topic/530363
To suppress this report, I suggest the use of
TSAN_OPTIONS='ignore_uninstrumented_modules=1'.
With this patch, a runtime warning is provided in case an OpenMP application
is built with Tsan and executed without this Tsan-option.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D70412
Summary:
[nfc][libomptarget] Refactor nxptx/target_impl.cu
Use __kmpc_impl_atomic_add instead of atomicAdd to match the rest of the file.
Alternatively, target_impl.cu could use the cuda functions directly. Using a mixture in this
file was an oversight, happy to resolve in either direction.
Removed some comments that look outdated.
Call __kmpc_impl_unset_lock directly to avoid a redundant diagnostic and remove an implict
dependency on interface.h.
Reviewers: ABataev, grokos, jdoerfert
Reviewed By: jdoerfert
Subscribers: jfb, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D72719
Summary:
[nfc][libomptarget] Refactor amdgcn target_impl
Removes references to internal libraries from the header
Standardises on C++ mangling for all the target_impl functions
Update comment block
clang-format
Move some functions into a new target_impl.hip source file
This lays the groundwork for implementing the remaining unresolved
symbols in the target_impl.hip source.
Reviewers: jdoerfert, grokos, ABataev, ronlieb
Reviewed By: jdoerfert
Subscribers: jvesely, mgorny, jfb, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D72712
The OpenMP runtime is not instrumented, so entering the runtime leaves no hint
on the source line of the pragma on ThreadSanitizer's function stack.
This patch adds function entry/exit annotations for OpenMP parallel regions,
and synchronization regions (barrier, taskwait, taskgroup).
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D70408
If the openmp project is built standalone, the test compiler is feature tested for an available -fsanitize=thread flag.
If the openmp project is built as part of llvm, the target tsan is needed to test archer.
An additional line (requires tsan) was introduced to the tests, this patch updates the line numbers for the race.
Follow-up for 77ad98c
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D71914
Summary:
If the dynamically loaded module has been compiled with -fopenmp-targets
and has no target regions, it has empty target descriptor. It leads to a
crash at the runtime if another module has at least one target region
and at least one entry in its descriptor. The runtime library is unable
to load the empty binary descriptor and terminates the execution.
Caused by a clang-offload-wrapper.
Reviewers: grokos, jdoerfert
Subscribers: caomhin, kkwli0, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D72472
Including two tests
These callbacks were added late to the 5.0 specification, an implementation is missing.
Reviewed By: jdoerfert
Differential Review: https://reviews.llvm.org/D70395
Summary:
[libomptarget][nfc] Introduce atomic wrapper function
Wraps atomic functions in a template prefixed __kmpc_atomic that
dispatches to cuda or hip atomic functions. Intended to be easily extended
to dispatch to OpenCL or C++ atomics for a third target.
Reviewers: ABataev, jdoerfert, grokos
Reviewed By: jdoerfert
Subscribers: Anastasia, jvesely, mgrang, dexonsmith, llvm-commits, mgorny, jfb, openmp-commits
Tags: #openmp, #llvm
Differential Revision: https://reviews.llvm.org/D71404
Summary:
[libomptarget][nfc] Extract function from data_sharing, move to common
Finding the first active thread in the warp is different on nvptx and amdgcn,
mostly due to warp size and the desire for efficiency.
Reviewers: ABataev, jdoerfert, grokos
Reviewed By: jdoerfert
Subscribers: jvesely, mgorny, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D71643
Summary:
[libomptarget][nfc] Move three files under common, build them for amdgcn
Change to reduction.cu to remove two dead includes, otherwise no code change.
Reviewers: jdoerfert, ABataev, grokos
Reviewed By: jdoerfert
Subscribers: jvesely, mgorny, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D71601
Summary:
[libomptarget][nfc] Move omp locks under target_impl
These are likely to be target specific, even down to the lock_t which is
correspondingly moved out of interface.h. The alternative is to include
interface.h in target_impl which substantiatially increases the scope of
those symbols.
The current nvptx implementation deadlocks on amdgcn. The preferred
implementation for that arch is still under discussion - this change
leaves declarations in target_impl.
The functions could be inline for nvptx. I'd prefer to keep the internals
hidden in the target_impl translation unit, but will add the (possibly renamed)
macros to target_impl.h if preferred.
Reviewers: ABataev, jdoerfert, grokos
Reviewed By: jdoerfert
Subscribers: jvesely, mgorny, jfb, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D71574
Summary:
[libomptarget][nfc] Wrap cuda min() in target_impl
nvptx forwards to cuda min, amdgcn implements directly.
Sufficient to build parallel.cu for amdgcn, added to CMakeLists.
All call sites are homogenous except one that passes a uint32_t and an
int32_t. This could be smoothed over by taking two type parameters
and some care over the return type, but overall I think the inline
<uint32_t> calling attention to what was an implicit sign conversion
is cleaner.
Reviewers: ABataev, jdoerfert
Reviewed By: jdoerfert
Subscribers: jvesely, mgorny, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D71580
Summary:
This reverts commit dd8a7fcdd7.
Alexey reports undefined symbols for the new inline functions defined in target_impl.h
This does not reproduce for me for nvptx, or amdgcn, under release or debug builds.
I believe the patch is fine, based on:
- the semantics of an inline function in C++ (the cuda INLINE functions end
up as linkonce_odr in IR), which are only legal to drop if they have no uses
- the code generated from a debug build of clang 9 does not show these undef symbols
- the tests pass
- the code is trivial
To progress from here I either need:
- A tie break - someone to play the role of CI in determining whether the patch works
- Alexey to provide sufficient information about his build for me to reproduce the failure
- Alexey to debug why the symbols are disappearing for him and report back
Reviewers: ABataev, jdoerfert, grokos
Subscribers: jvesely, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D71502
Summary:
[libomptarget] Build most of common/src for amdgcn
Excluding parallel.cu, which uses an integer min() from cuda,
Excluding support.cu, which calls malloc that is not yet available for amdgcn
Reviewers: jdoerfert, ABataev, grokos
Reviewed By: jdoerfert
Subscribers: gregrodgers, ronlieb, jvesely, mgorny, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D71446
Summary:
[libomptarget][nfc] Add declarations of atomic functions for amdgcn
This enables building more source for amdgcn. The functions are usually available
in a hip runtime header, but are duplicated here to decouple the implementation
Reviewers: jdoerfert, ABataev, grokos
Reviewed By: jdoerfert
Subscribers: jvesely, mgorny, jfb, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D71412
Summary:
[libomptarget][nfc] Move cuda threadfence functions behind kmpc_impl
Part of building code under common/ without requiring a cuda compiler
Reviewers: ABataev, jdoerfert, grokos
Reviewed By: ABataev
Subscribers: jvesely, jfb, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D71102
Summary:
[libomptarget][nfc] Move omptarget-nvptx under common
Almost all files depend on require omptarget-nvptx, which no longer
contains any obviously architecture dependent code. Moving it under
common unblocks task/loop for amdgcn, and allows moving other code.
At some point there should probably be a widespread symbol renaming to
replace the nvptx string. I'd prefer to get things working first.
Building this (and task.cu, loop.cu) without a cuda library requires
some more refactoring, e.g. wrap threadfence(), use DEVICE macro more
consistently. Patches for that are orthogonal and will be posted shortly.
Reviewers: jdoerfert, ABataev, grokos
Reviewed By: ABataev
Subscribers: mgorny, fedor.sergeev, jfb, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D71073
Summary:
[libomptarget] Build a minimal deviceRTL for amdgcn
Repeat of D70414, with an include path fixed. Diff for sanity checking.
The CMakeLists.txt file is functionally identical to the one used in the aomp fork.
Whitespace changes were made based on nvptx/CMakeLists.txt, plus the
copyright notice updated to match (Greg was the original author so would
like his sign off on that here).
This change will build a small subset of the deviceRTL if an appropriate toolchain is
available, e.g. a local install of rocm. Support.h is moved from nvptx as a dependency
of debug.h.
Reviewers: ABataev, jdoerfert
Reviewed By: ABataev
Subscribers: jvesely, mgorny, jfb, openmp-commits, jdoerfert
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D70971
Summary:
[libomptarget] Build a minimal deviceRTL for amdgcn
The CMakeLists.txt file is functionally identical to the one used in the aomp fork.
Whitespace changes were made based on nvptx/CMakeLists.txt, plus the
copyright notice updated to match (Greg was the original author so would
like his sign off on that here).
This change will build a small subset of the deviceRTL if an appropriate toolchain is
available, e.g. a local install of rocm. Support.h is moved from nvptx as a dependency
of debug.h.
Reviewers: jdoerfert, ABataev, grokos, ronlieb, gregrodgers
Reviewed By: jdoerfert
Subscribers: jfb, Hahnfeld, jvesely, mgorny, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D70414
Summary:
"make check-all" or "make check-libomptarget" would attempt to run offloading
tests before the offload plugins are built. This patch corrects that by adding
dependencies to the libomptarget CMake rules.
Reviewers: jdoerfert
Subscribers: mgorny, guansong, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D70803
Summary:
The termination function duplicated the functionality of the
__attribute((destructor))-annotated function __kmp_internal_end_fini,
and we have no indication that this doesn't work.
The function might cause issues with link-time optimization turned on:
until very recently, none of the usual linkers was reporting functions
named in -Wl,-fini as used to the LTO plugin, so it might be dropped.
If the function is dropped, -Wl,-fini=__kmp_internal_end_fini doesn't
do what we want: with ld.bfd and lld it drops the FINI attribute from
.dynamic and with gold we get FINI = 0x0, which leads to a crash on
cleanup. This can be reproduced by building with
-DLLVM_ENABLE_PROJECTS="clang;openmp" \
-DLLVM_ENABLE_LTO=Thin \
-DLLVM_USE_LINKER=gold
The issue in lld has been fixed in f95273f75a, but gold remains without
fix so far.
Fixes PR43927.
Reviewers: JonChesterfield, jdoerfert, AndreyChurbanov
Reviewed By: AndreyChurbanov
Differential Revision: https://reviews.llvm.org/D69927
Summary:
[libomptarget][nfc] Move some source into common from nvptx
Moves some source that compiles cleanly under amdgcn into a common subdirectory
Includes some non-trivial files and some headers. Keeps the cuda file extension.
The build systems for different architectures seem unlikely to have much in
common. The idea is therefore to set include paths such that files under
common/src compile as if they were under arch/src as the mechanism for sharing.
In particular, files under common/src need to be able to include target_impl.h.
The corresponding -Icommon is left out in favour of explicit includes on the
basis that the it makes it clearer which files under common are used by a given
architecture.
Reviewers: jdoerfert, ABataev, grokos
Reviewed By: ABataev
Subscribers: jfb, mgorny, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D70328
The tool provides TSAN annotations for OpenMP synchronization. The tool
is activated if no other OMPT tool is loaded.
The tool detects whether the application was built with TSan and rejects
activation according to the OMPT protocol if there is no TSan-rt.
Differential Revision: https://reviews.llvm.org/D45890
Summary:
[libomptarget][nfc] Use cuda variable wrappers from support.h
Reimplementation of D69693, after the revert of D69885
Use the wrappers in support.h for cuda builtin variables at all call sites.
Localises use of cuda and removes WARPSIZE==32 assumption in debug.h.
Reviewers: ABataev, jdoerfert, grokos
Reviewed By: jdoerfert
Subscribers: openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D70186
Summary:
[libomptarget] Move supporti.h to support.cu
Reimplementation of D69652, without the unity build and refactors.
Will need a clean build of libomptarget as the cmakelists changed.
Reviewers: ABataev, jdoerfert
Reviewed By: jdoerfert
Subscribers: mgorny, jfb, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D70131
Summary:
[libomptarget] Revert all improvements to support
The change to unity build for nvcc has broken the build for some developers.
This patch reverts to a known-working state.
There has been some confusion over exactly how the build broke. I think we
have reached a common understanding that the disappearing symbols are from
the bitcode library built by clang. The static archive built by nvcc may show the
same problem. Some of the confusion arose from building the deviceRTL twice
and using one or the other library based on various environmental factors.
I'm pretty sure the problem is clang expanding `__forceinline__` into both `__inline__`
and `attribute(("always_inline"))`. The `__inline__` attribute resolves to linkonce_odr
which is not safe for exporting symbols from translation units.
"always_inline" is the desired semantic for small functions defined in one translation
unit that are intended to be inlined at link time. "inline" is not.
This therefore reintroduces the dependency hazard of supporti.h and some code
duplication, and blocks progress separating deviceRTL into reusable components.
See also D69857, D69859 for attempts at a fix instead of a revert.
Reviewers: ABataev, jdoerfert, grokos, ikitayama, tianshilei1992
Reviewed By: ABataev
Subscribers: mgorny, jfb, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D69885
Summary:
[libomptarget] Implement target_impl for amdgcn
Smallest atomic addition for a new target. Implements enough of the amdgcn
specific code that some of the source files under nvptx/src could be compiled,
without modification, to run on amdgcn.
This foreshadows a work in progress patch to move said source out of nvptx/src.
Patch based on fork at https://github.com/ROCm-Developer-Tools/llvm-project
Reviewers: ABataev, jdoerfert, grokos, ronlieb
Subscribers: jvesely, jfb, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D69718
Summary:
[nfc][omptarget] Use builtin var abstraction. Second pass at D69476
Use the wrappers in support.h for cuda builtin variables at all call sites.
Localises use of cuda and removes WARPSIZE==32 assumption in debug.h.
Reviewers: ABataev, jdoerfert, grokos
Reviewed By: jdoerfert
Subscribers: openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D69693
Summary:
[nfc][libomptarget] Reorganise support header
All functions defined in support implementation are now declared in support.h
Reordered functions in support implementation to match the sequence in support.h
Added include guards to support.h
Added #include interface to support.h to provide kmp_Ident declaration
Move supporti.h to support.cu and s/INLINE/EXTERN/g
Add remaining includes to support.cu
A minor side effect is to change the name mangling of the support functions to
extern "C". If this matters another macro along the lines of INLINE/EXTERN
can be added - perhaps DEVICE as that's the obvious implementation.
Reviewers: jdoerfert, ABataev, grokos
Reviewed By: jdoerfert
Subscribers: mgorny, jfb, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D69652
Summary:
[libomptarget] Change nvcc compilation to use a unity build
This allows nvcc to inline functions between what would otherwise be distinct
translation units, which in turn removes any runtime cost from implementing
functions in source files (as opposed to inline in headers).
This will then allow the circular dependencies in deviceRTL to be readily
broken and individual components more easily shared between architectures.
Reviewers: ABataev, jdoerfert, grokos, RaviNarayanaswamy, hfinkel, ronlieb, gregrodgers
Reviewed By: jdoerfert
Subscribers: mgorny, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D69489
Summary:
[libomptarget] Always call malloc, free via SafeMalloc, SafeFree wrapper
NFC for release, adds some verbosity to debug printing. Motivation is to provide
one place where local modifications can be made to the behaviour of all heap
allocation or deallocation while debugging.
Reviewers: jdoerfert, ABataev, grokos
Reviewed By: ABataev
Subscribers: openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D69492
Summary:
[nfc][libomptarget] Decrease coupling between files
debug.h used the symbol omptarget_device_environment so implicitly required
an include of omptarget-nvptx.h to compile. Similarly interface.h uses size_t.
Moving this declaration to a new header means cancel, critical can now build
without omptarget-nvptx.h. After this change, debug.h, cancel.cu, critical.cu
could move under a common source directory.
Reviewers: ABataev, jdoerfert, grokos
Subscribers: openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D69473
Summary:
[nfc][libomptarget] Inline option into target_impl
Subset of D69423. The macros that were in option.h are all target dependent.
Inlining the header simplifies the dependency graph when looking to move code
into a common subdir.
Reviewers: ABataev, jdoerfert, grokos
Subscribers: openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D69472
Summary:
[NFC][libomptarget] move remaining device specific code out of omptarget-nvptx.h
Strictly there is one remaining difference wrt amdgcn - parallelLevel is
volatile qualified on amdgcn and not on nvptx. Determining whether this is
correct - and how to represent the different semantics of 'volatile' under
various conditions - is beyond the scope of this code motion patch.
Reviewers: ABataev, jdoerfert, grokos
Subscribers: openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D69424
Details:
- nconflicts field initialized;
- formatting fix (moved declaration out of the long line);
- count conflicts in new hash as opposed to old one.
Differential Revision: https://reviews.llvm.org/D68036
Summary:
[libomptarget][nfc] Make interface.h target independent
Move interface.h under a top level include directory.
Remove #includes to avoid the interface depending on the implementation.
Reviewers: ABataev, jdoerfert, grokos, ronlieb, RaviNarayanaswamy
Reviewed By: jdoerfert
Subscribers: mgorny, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D68615
llvm-svn: 374919
Summary:
[libomptarget][nfc] Update remaining uint32 to use lanemask_t
Update a few functions in the API to use lanemask_t instead of i32. NFC for
nvptx. Also update the ActiveThreads type in DataSharingStateTy.
This removes a lot of #ifdef from the downsteam amdgcn implementation.
Reviewers: ABataev, jdoerfert, grokos, ronlieb, RaviNarayanaswamy
Subscribers: openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D68513
llvm-svn: 373806
/proc unless Linux layer compatibility is activated for CentOS is activated is not present
thus relying on a more native for checking the address.
Reviewers: Hahnfeld, kongyl, jdoerfert, jlpeyton, AndreyChurbanov, emaster, dim
Reviewed By: Hahnfeld
Differential Revision: https://reviews.llvm.org/D67326
llvm-svn: 373152
Linker automatically provides __start_<section name> and __stop_<section name> symbols to satisfy unresolved references if <section name> is representable as a C identifier (see https://sourceware.org/binutils/docs/ld/Input-Section-Example.html for details). These symbols indicate the start address and end address of the output section respectively. Therefore, renaming OpenMP offload entries section name from ".omp.offloading_entries" to "omp_offloading_entries" to use this feature.
This is the first part of the patch for eliminating OpenMP linker script (please see https://reviews.llvm.org/D64943).
Differential Revision: https://reviews.llvm.org/D68070
llvm-svn: 373118
There's no need to initialize variables with static storage duration
because they're implicitly initialized to zero. See
https://en.cppreference.com/w/c/language/initialization#Implicit_initialization
I think that's already relied upon because the supplied 0 only sets
'kmp_time_global_t g_time;' in 'struct kmp_base_global'. The other fields
are not set in the code, but implicitly initialized by the compiler.
Differential Revision: https://reviews.llvm.org/D66292
llvm-svn: 370943
Summary:
In non-SPMD mode we may end up with the divergent threads when trying to
increment/decrement parallel level counter. It may lead to incorrect
calculations of the parallel level and wrong results when threads are
divergent. We need to reconverge the threads before trying to modify the
parallel level counter.
Reviewers: grokos, jdoerfert
Subscribers: guansong, openmp-commits, caomhin, kkwli0
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D66802
llvm-svn: 370803
Summary:
Use target_impl functions to replace more inline asm
Follow on from D65836. Removes remaining asm shuffles and lanemask accessors
Also changes the types of target_impl bitwise functions to unsigned.
Reviewers: jdoerfert, ABataev, grokos, Hahnfeld, gregrodgers, ronlieb, hfinkel
Subscribers: openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D66809
llvm-svn: 370216
Summary:
[libomptarget] Refactor syncthreads macro to inline function
See also abandoned D66846, split into this diff and others.
Rev 2 of D66855
Reviewers: jdoerfert, ABataev, grokos, ronlieb, gregrodgers
Subscribers: openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D66861
llvm-svn: 370210
Summary:
[libomptarget] Refactor syncwarp macro to inline function
See also abandoned D66846, split into this diff and others.
Reviewers: jdoerfert, ABataev, grokos, ronlieb, gregrodgers
Subscribers: openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D66857
llvm-svn: 370149
Summary:
[libomptarget] Refactor shfl_down_sync macro to inline function
See also abandoned D66846, split into this diff and others.
Reviewers: jdoerfert, ABataev, grokos, ronlieb, gregrodgers
Subscribers: openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D66853
llvm-svn: 370146
Summary:
[libomptarget] Refactor shfl_sync macro to inline function
See also abandoned D66846, split into this diff and others.
Reviewers: jdoerfert, ABataev, grokos, ronlieb, gregrodgers
Subscribers: openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D66852
llvm-svn: 370144
Summary:
Added function void __kmpc_syncwarp(int32_t) to expose it to the
compiler. It is required to fix the problem with the critical regions in
Cuda9.0+. We cannot use barrier in the critical region, but still need
to reconverge the threads in the warp after. This function allows to do
this.
Reviewers: grokos, jdoerfert
Subscribers: guansong, openmp-commits, kkwli0, caomhin
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D66672
llvm-svn: 369933
Summary:
In Cuda 9.0 it is not guaranteed that threads in the warps are
convergent. We need to use __syncwarp() function to reconverge
the threads and to guarantee the memory ordering among threads in the
warps.
This is the first patch to fix the problem with the test
libomptarget/deviceRTLs/nvptx/src/sync.cu on Cuda9+.
This patch just replaces calls to __shfl_sync() function with the call
of __syncwarp() function where we need to reconverge the threads when we
try to modify the value of the parallel level counter.
Reviewers: grokos
Subscribers: guansong, jfb, jdoerfert, caomhin, kkwli0, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D65013
llvm-svn: 369796
This patch fixes https://bugs.llvm.org/show_bug.cgi?id=42906, via adding
adjustment of number of threads on enter to the teams construct on host
according to user settings. This allows to pass checks and avoid assertions
at time of team of threads creation.
Patch by Andrey Churbanov
Differential Revision: https://reviews.llvm.org/D66351
llvm-svn: 369430
Fix last warned location in ittnotify_static.cpp using the defined
macro KMP_FALLTHROUGH().
Differential Revision: https://reviews.llvm.org/D65871
llvm-svn: 369003
The variables in kmp_lock.cpp are really arrays of function pointers
that return void or int, not pointers to functions that return void*
or int*. The other changes are only cosmetic.
Differential Revision: https://reviews.llvm.org/D65870
llvm-svn: 369002
The implementation status can only be one of
ompt_event_UNIMPLEMENTED = ompt_set_never = 1
ompt_event_MAY_ALWAYS = ompt_set_always = 5
In both cases, the condition was already true, so just remove
the check.
Differential Revision: https://reviews.llvm.org/D65869
llvm-svn: 369001
Instead, maintain a list of disabled options to still build libomp and
libomptarget without warnings. This includes -Wno-error and -Wno-pedantic
to silence warnings that LLVM enables when building in-tree.
I tested the following compilers:
* Clang 6.0, 7.0, 8.0
* GCC 4.8.5 (CentOS 7), GCC 6, 7, 8, 9
* Intel Compiler 16, 17, 18, 19
RFC thread on openmp-dev mailing list:
http://lists.llvm.org/pipermail/openmp-dev/2019-August/002668.html
Differential Revision: https://reviews.llvm.org/D65867
llvm-svn: 368999
Summary:
[libomptarget] Factor architecture dependent code out of loop.cu
Related to the patch series starting D64217. Added subscribers to said series as reviewers. This effort is smaller in scope.
This patch factors out just enough architecture dependent code from loop.cu to allow the same source to be used with amdgcn, given a different target_impl.h. Testing is that the same bitcode (modulo variable names) is generated for libomptarget before and after the refactor, for nvptx and the out of tree amdgcn.
Reviewers: jdoerfert, ABataev, bollu, jfb, tra, grokos, Hahnfeld, guansong, xtian, gregrodgers, ronlieb, hfinkel, gtbercea, guraypp, arpith-jacob
Reviewed By: jdoerfert, ABataev
Subscribers: dexonsmith, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D65836
llvm-svn: 368751
This patch fixes problem raised in post-review comments of the
https://reviews.llvm.org/D65285. Developers of ittnotify confirmed
that dll_path_ptr field of the __itt_global structure is never used
by ittnotify library, so it is safe to remove the dll_path array.
Differential Revision: https://reviews.llvm.org/D65885
llvm-svn: 368559
Summary:
This patch adds support for the close map modifier.
The close map modifier will overwrite the unified shared memory requirement and create a device copy of the data.
Reviewers: ABataev, Hahnfeld, caomhin, grokos, jdoerfert, AlexEichenberger
Reviewed By: Hahnfeld, AlexEichenberger
Subscribers: guansong, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D65340
llvm-svn: 368488
We have one global RTLs.RequiresFlags, I don't see a need to make a
copy per device that the runtime manages. This was problematic anyway
because the copy happened during the first __tgt_register_lib(). This
made it impossible to call __tgt_register_requires() from normal user
funtions for testing.
Hence, this change also fixes unified_shared_memory/shared_update.c for
older versions of Clang that don't call __tgt_register_requires() before
__tgt_register_lib().
Differential Revision: https://reviews.llvm.org/D66019
llvm-svn: 368465
Summary:
This patch adds support for using unified memory in the case of regular maps that happen when a target region is offloaded to the device.
For cases where only a single version of the data is required then the host address can be used. When variables need to be privatized in any way or globalized, then the copy to the device is still required for correctness.
Reviewers: ABataev, jdoerfert, Hahnfeld, AlexEichenberger, caomhin, grokos
Reviewed By: Hahnfeld
Subscribers: mgorny, guansong, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D65001
llvm-svn: 368192
Summary:
[libomptarget] Use forceinline. Necessary for nvcc to inline small functions within the bitcode library
Suggested in D65836
Reviewers: ABataev, jdoerfert, grokos, gregrodgers
Subscribers: openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D65876
llvm-svn: 368177
New OMPT tests with teams construct should be disabled for GCC as it
emits code with a GOMP entry not supported in the LLVM runtime.
Differential Revision: https://reviews.llvm.org/D65757
llvm-svn: 367939
Ensures that CUDA fail reasons (such as "No CUDA-capable device detected")
are printed together with libomptarget's debug message
(e.g. "Error when setting CUDA context"). Previously, the former was
printed only in CMAKE_BUILD_TYPE=Debug builds while the latter was
enabled by LIBOMPTARGET_ENABLE_DEBUG.
With this change, also only call cuGetErrorString when the error will be
printed.
Suggested-by: Ye Luo <xw111luoye@gmail.com>
Differential Revision: https://reviews.llvm.org/D65687
llvm-svn: 367910
This patch implements the libomptarget runtime interface for OpenMP 5.0
declare mapper functions. The declare mapper functions generated by
Clang will call them to complete the mapping of members.
kmpc_mapper_num_components gets the current number of components for a
user-defined mapper; kmpc_push_mapper_component pushes back one
component for a user-defined mapper.
The design slides can be found at
https://github.com/lingda-li/public-sharing/blob/master/mapper_runtime_design.pptx
Patch by Lingda Li <lildmh@gmail.com>
Differential Revision: https://reviews.llvm.org/D60972
llvm-svn: 367772
All other files are already C++ and the build system has always
passed '-x c++' for C files, effectively compiling them as C++.
To stay warning free we need one fix in ittnotify_static.{c,cpp}:
The variable dll_path can be written to, so it must not be const.
GCC complained with -Wcast-qual and I think it's right.
Differential Revision: https://reviews.llvm.org/D65285
llvm-svn: 367343
Round the stack size to a multiple of the page size. Older versions of
Android (until KitKat) would fail pthread_attr_setstacksize with
EINVAL if the stack size was not a multiple of the page size.
Patch by Dan Albert <danalbert@google.com>.
Test: Build, copied into the NDK, passed openmp test on ICS.
Bug: https://github.com/android-ndk/ndk/issues/9
llvm-svn: 367070
Both Clang and GCC complained that they cannot initialize a return
object of type 'kmp_proc_bind_t' with an 'int'. While at it, also
fix a warning about missing parentheses thrown by Clang.
Differential Revision: https://reviews.llvm.org/D65284
llvm-svn: 367041
Summary:
According to the OpenMP standard, barrier operation must perform
implicit flush operation. Currently, if there is only one thread in the
team, barrier does not flush the memory. Patch fixes this problem.
Reviewers: grokos, gtbercea, kkwli0
Subscribers: guansong, jdoerfert, openmp-commits, caomhin
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D62398
llvm-svn: 367024
This is a port of libomp for the RISC-V 64-bit Linux target.
We have tested this port on a HiFive Unleashed development board
using a downstream LLVM that has support for the missing bits in
upstream. As of now, all tests are passing, including OMPT.
Patch by Ferran Pallarès!
Differential Revision: https://reviews.llvm.org/D59880
llvm-svn: 367021
If the first target region in a program calls the push_tripcount
function, libomptarget didn't handle the offload policy correctly.
This could lead to unexpected error messages as seen in
http://lists.llvm.org/pipermail/openmp-dev/2019-June/002561.html
To solve this, add a check calling IsOffloadDisabled() as all other
entry points already do. If this method returns false, libomptarget
is effectively disabled.
Differential Revision: https://reviews.llvm.org/D64626
llvm-svn: 366810
This is done at call-site and does not need to be handled in
__kmp_invoke_microtask. It was already absent from the x86
and x86_64 assembly, this patch removes it from the generic
implementation in z_Linux_util.cpp and adds documentation for
AArch64 and PPC64 that it's actually not needed. I can't test
on these architectures, so I don't want to change the code just
because it looks right :)
While at it, rename some variables for consistency and add a
check in test/ompt/parallel/normal.c that the pointer was reset
before entering the barrier.
Differential Revision: https://reviews.llvm.org/D64442
llvm-svn: 366721
Remove loopTripCnt from threaded device stack after consuming it.
Added a libomptarget DP message to aid in future debugging and to
validate the added testcase, which only runs in Debug build.
Differential Revision: https://reviews.llvm.org/D64808
llvm-svn: 366349
This is a follow up patch to D64534 (r365963) which removed all OMP
spec versioning within the OpenMP runtime codebase. This patch removes
REQUIRES: openmp-x.y lines from lit tests.
llvm-svn: 366341
This leads to problems when compiling C++ code with libc++ for Nvidia GPUs
because Clang now uses wrappers for math functions that might include
C++ templates not allowed in 'extern "C"'.
Differentiel Revision: https://reviews.llvm.org/D64625
llvm-svn: 366229
Summary:
We used CUDART_VERSION macro to check for the installed cuda version
but this macro is defined in cuda_runtime_api.h, which is not used by
project. Better to use CUDA_VERSION macro, which is defined in cuda.h.
Also, added the check if this macro is defined. If macro is undefined,
there is something wrong with the cuda configuration and we should not
continue the compilation.
This also fixes problems with runtime building in cuda 10+.
Reviewers: grokos
Subscribers: guansong, jdoerfert, caomhin, kkwli0, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D64648
llvm-svn: 366224
Summary:
We used to call __kmpc_omp_taskwait function with global threadid set to
0. It may crash the application at the runtime if the thread executing
target region is not a master thread.
Reviewers: grokos, kkwli0
Subscribers: guansong, jdoerfert, caomhin, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D64571
llvm-svn: 366220
Remove all older OMP spec versioning from the runtime and build system.
Patch by Terry Wilmarth
Differential Revision: https://reviews.llvm.org/D64534
llvm-svn: 365963
These entry points are never called by Clang trunk nor clang-ykt. If
XL doesn't use them either, they can finally go away.
Differential Revision: https://reviews.llvm.org/D52700
llvm-svn: 365817
Summary:
__kmpc_push_tripcount function is not thread safe and may lead to data
race when the target regions are executed in parallel threads. The patch
makes loopTripCnt counter thread aware and stores the tripcount value
per thread in the map. Access to map is guarded by mutex to prevent
data race in the map itself.
Test is for NVPTX target because it does not work correctly on the
host. Seems to me, there is a problem in libomp with target regions in
the parallel threads.
Reviewers: grokos
Subscribers: guansong, jfb, jdoerfert, openmp-commits, kkwli0, caomhin
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D64080
llvm-svn: 365332
Summary:
According to the OpenMP standard, flush makes a thread’s temporary view of memory consistent with memory and enforces an order on the memory operations of the variables explicitly specified or implied.
According to the Cuda toolkit documentation (https://docs.nvidia.com/cuda/archive/8.0/cuda-c-programming-guide/index.html#memory-fence-functions), __threadfence() functions provides required functionality.
__threadfence_system() also provides required functionality, but it also
includes some extra functionality, like synchronization of page-locked
host memory, synchronization for the host, etc. It is not required per
the standard and we can use more relaxed version of memory fence
operation.
Reviewers: grokos, gtbercea, kkwli0
Subscribers: guansong, jfb, jdoerfert, openmp-commits, caomhin
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D62397
llvm-svn: 364572
Bug reported in https://bugs.llvm.org/show_bug.cgi?id=42269.
Freeing of the contention group (CG) stucture by master thread looks wrong,
because workers can leave the CG later on. Intead the freeing
is now done by the last thread leaving the CG.
Differential Revision: https://reviews.llvm.org/D63599
llvm-svn: 364456
Summary:
This patch adds support for handling variables under the:
```
#pragma omp declare target to()
```
clause when the
```
#pragma omp requires unified_shared_memory
```
is used.
The address of the host variable is copied into the device pointer just like for the declare target link case.
Reviewers: ABataev, caomhin, grokos, AlexEichenberger
Reviewed By: grokos
Subscribers: jcownie, guansong, jdoerfert, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D63106
llvm-svn: 363825
Summary:
The problems with __syncthreads() were fixed in clang >= 9.0 and the
original __syncthreads() can be used instead of the ptx instruction.
Reviewers: grokos
Subscribers: guansong, jdoerfert, openmp-commits, kkwli0, caomhin
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D63515
llvm-svn: 363807
Summary: This patch enables the usage of a host variable on the device for declare target link variables when unified memory is available.
Reviewers: ABataev, caomhin, grokos
Reviewed By: grokos
Subscribers: Hahnfeld, guansong, jdoerfert, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D60884
llvm-svn: 362505
Made type of depth of hwloc object to correapond with
change from unsigned in hwloc 1,x to int in hwloc 2.x.
This eliminates the warning on signed-unsigned comparison.
Differential Revision: https://reviews.llvm.org/D62332
llvm-svn: 362401
Current parsing allows trailing string after the permitted value,
MANDATORY|DISABLED|DEFAULT -- e.g., "mandatorynot" is also recognized
as "MANDATORY". Such cases should be recognized as incorrect/unknown
value.
Differential Revision: https://reviews.llvm.org/D62431
llvm-svn: 362125
This change adds checks before dereferencing a pointer returned from a
function.
Differential Revision: https://reviews.llvm.org/D62224
llvm-svn: 362111
The omp_taskloop_num_tasks and omp_taskwait have deadlooped
on the NetBSD buildbot previously, practically hanging the host running
it. Disable them until we can find a good solution, or make the kernel
less fragile.
llvm-svn: 361825
Summary:
Parallel level counter should be volatile to prevent some dangerous
optimiations by the ptxas. Otherwise, ptxas optimizations lead to
undefined behaviour in some cases.
Also, use __threadfence() for #pragma omp flush and if the barrier
should not be used (we have only one thread in the team), still perform
flush operation since the standard requires implicit flush when
executing barriers.
Reviewers: gtbercea, kkwli0, grokos
Subscribers: guansong, jfb, jdoerfert, openmp-commits, caomhin
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D62199
llvm-svn: 361421
This change adds implementation to ompt_finalize_tool() and
ompt_get_task_memory().
Patch by Hansang Bae
Differential Revision: https://reviews.llvm.org/D61657
llvm-svn: 361309
Summary:
Target link variables are currently implemented by creating a copy of the variables on the device side and unified memory never gets exploited.
When the prgram uses the:
```
#pragma omp requires unified_shared_memory
```
directive in conjunction with a declare target link, the linked variable is no longer allocated on the device and the host version is used instead.
This behavior is overridden by performing an explicit mapping.
A Clang side patch is required.
Reviewers: ABataev, AlexEichenberger, grokos, Hahnfeld
Reviewed By: AlexEichenberger, grokos, Hahnfeld
Subscribers: Hahnfeld, jfb, guansong, jdoerfert, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D60223
llvm-svn: 361294
https://reviews.llvm.org/D58454 did not fix the problem for a typical use
case of building LLVM with gcc or icc and then testing with the newly built
clang compiler.
The compilers do not agree on how to extend a 32-bit pointer to uint64, so
make the pointer unsigned first, before adjusting the size.
Patch by Joachim Protze
Differential Revision: https://reviews.llvm.org/D58506
llvm-svn: 361158
OpenMP 5.0 says that the callback for the events initial-task-begin and
initial-task-end has to be ompt_callback_implicit_task.
Patch by Tim Cramer
Differential Revision: https://reviews.llvm.org/D58776
llvm-svn: 361157
Added synchronization for possible concurrent initialization of mutexes
by multiple threads. The need of synchronization caused by commit r357927
which added the use of mutexes at threads movement to/from common pool
(earlier the mutexes were used only at suspend/resume).
Patch by Johnny Peyton.
Differential Revision: https://reviews.llvm.org/D61995
llvm-svn: 360919
Currently cores within package that share the same L2 cache are grouped together.
The current logic behind this assumes that the L2 cache is always at deeper
(or the same) level than the package itself. In case when L2 cache is common
for all packages (and the packages are at deeper level than L2 cache) the whole of
the further topology discovery fails to find any computational units resulting in
following assertion:
Assertion failure at kmp_affinity.cpp(715): nActiveThreads == __kmp_avail_proc.
OMP: Error #13: Assertion failure at kmp_affinity.cpp(715).
This patch adds a bit of a logic that prevents such situation from occurring.
Differential Revision: https://reviews.llvm.org/D61796
llvm-svn: 360890
Removed unconditional and unsafe decrement of counter
of active threads in pool at shutdown time.
Differential Revision: https://reviews.llvm.org/D61944
llvm-svn: 360784
The implementation should be done by compiler, user can only declare
objects of this type and use them in OpenMP directives.
Differential Revision: https://reviews.llvm.org/D61860
llvm-svn: 360774
The code is currently using the ambiguous instruction
"sub sp, sp, w9, lsl #4". The ARM reference manual says this isn't
valid, and it's not clear whether it's supposed to mean uxtw or uxtx.
It doesn't matter which instruction we use here, since the high
bits of the operand are zero anyway, so I arbitrarily choose uxtw, to
preserve the register name.
See https://reviews.llvm.org/D60840 for the LLVM patch.
Differential Revision: https://reviews.llvm.org/D61770
llvm-svn: 360711
Changed file extension of the destination of the copy of libomp.lib
(it was mistakely .dll, now it is .lib) in installation on Windows.
Differential Revision: https://reviews.llvm.org/D61673
llvm-svn: 360595
Summary:
Patch improves performance of the full runtime mode by moving
threads limit counter to the shared memory. It also allows to save
global memory.
Reviewers: grokos, kkwli0, gtbercea
Subscribers: guansong, jdoerfert, openmp-commits, caomhin
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D61801
llvm-svn: 360584
Summary:
Patch improves performance of the full runtime mode by moving
number-of-threads counter to the shared memory. It also allows to save
global memory.
Reviewers: grokos, gtbercea, kkwli0
Subscribers: guansong, jfb, jdoerfert, openmp-commits, caomhin
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D61785
llvm-svn: 360457
This patch provides workaround to allow gfortran to compile the
OpenMP Fortran modules.
From the gfortran manual:
https://gcc.gnu.org/onlinedocs/gcc-9.1.0/gfortran/BOZ-literal-constants.html
"Note that initializing an INTEGER variable with a statement such as
DATA i/Z'FFFFFFFF'/ will give an integer overflow error rather than the desired
result of -1 when i is a 32-bit integer on a system that supports 64-bit
integers. The -fno-range-check option can be used as a workaround for legacy
code that initializes integers in this manner."
Bug filed: https://bugs.llvm.org/show_bug.cgi?id=41755
Differential Revision: https://reviews.llvm.org/D61603
llvm-svn: 360299
Summary:
To be able to successfully build OpenMP on FreeBSD/i386, which still
uses i486 as its default processor, I had to provide wrappers for the
`__kmp_load_mxcsr` and `__kmp_store_mxcsr` functions.
If the compiler signals that SSE is not available, loading and storing
mxcsr does not make sense anway, so in that case the inline functions
are empty. This gives the minimum amount of code churn.
See also https://svnweb.freebsd.org/changeset/base/345283
Reviewers: emaste, jlpeyton, Hahnfeld
Reviewed By: jlpeyton
Subscribers: hfinkel, krytarowski, jdoerfert, openmp-commits, llvm-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D60916
llvm-svn: 360062
Summary:
Patch improves performance of the full runtime mode by moving
thread-limit counter to the shared memory. It also allows to save
global memory.
Reviewers: grokos, gtbercea, kkwli0
Subscribers: guansong, jdoerfert, caomhin, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D61526
llvm-svn: 359922
Summary:
Used parallelLevel[] counter to simplify and improve implementation of
the existing standard OpenMP functions. Functions are tested already in
several tests, the patch is NFC.
Reviewers: grokos, gtbercea, kkwli0
Subscribers: guansong, jdoerfert, caomhin, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D61459
llvm-svn: 359892
Summary:
Previously for the different purposes we need to get the active/common
parallel level and with full runtime we iterated over all the records to
calculate this level. Instead, we can used the warp-based parallel level
counters used in no-runtime mode.
Reviewers: grokos, gtbercea, kkwli0
Subscribers: guansong, jfb, jdoerfert, caomhin, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D61395
llvm-svn: 359822
Summary:
Function omp_get_thread_limit() in SPMD mode can return the maximum
available number of threads as a result.
Reviewers: grokos, gtbercea, kkwli0
Subscribers: guansong, jdoerfert, openmp-commits, caomhin
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D61378
llvm-svn: 359790
Summary:
To be able to successfully build OpenMP on 32-bit FreeBSD, such as
FreeBSD/i386, I first had to provide a few wrappers (see D60916), and
then add `KMP_OS_FREEBSD` to the list of defines checked for 32-bit
architectures in `kmp_runtime.cpp`.
I have successfully built libomp.so and ran a bunch of test programs on
FreeBSD/i386 with this.
See also https://svnweb.freebsd.org/changeset/base/345283
Reviewers: emaste, jlpeyton, Hahnfeld
Reviewed By: jlpeyton
Subscribers: krytarowski, guansong, jdoerfert, openmp-commits, llvm-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D60917
llvm-svn: 359716
Implemented task modifier in two versions - one without taking into account
omp_orig variable (the omp_orig still can be processed by compiler without help
of the library, but each reduction object will need separate initializer with
global access to omp_orig), another with omp_orig variable included into
interface (single initializer can be used for multiple reduction objects of
the same type). Second version can be used when the omp_orig is not globally
accessible, or to optimize code in case of multiple reduction objects
of the same type.
Patch by Andrey Churbanov
Differential Revision: https://reviews.llvm.org/D60976
llvm-svn: 359710
This patch adds:
* New omp_sched_monotonic flag to omp_sched_t which is handled within the runtime
* Parsing of monotonic/nonmonotonic in OMP_SCHEDULE
* Tests for the monotonic flag and envirable parsing
* Logic to force monotonic when hierarchical scheduling is used
Differential Revision: https://reviews.llvm.org/D60979
llvm-svn: 359601
Summary:
The parallelLevel counter must be on per-thread basis to fully support
L2+ parallelism, otherwise we may end up with undefined behavior.
Introduce the parallelLevel on per-warp basis using shared memory. It
allows to avoid the problems with the synchronization and allows fully
support L2+ parallelism in SPMD mode with no runtime.
Reviewers: gtbercea, grokos
Subscribers: guansong, jdoerfert, caomhin, kkwli0, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D60918
llvm-svn: 359341
Summary:
I ran into some issues after rOMP355687, where __atomic_fetch_add was
being used incorrectly on x86, and this turns out to be caused by the
following added conditionals:
```
#if defined(KMP_ARCH_MIPS)
```
The problem is, these macros are always defined, and are either 0 or 1
depending on the architecture. E.g. the correct way to test for MIPS
is:
```
#if KMP_ARCH_MIPS
```
Reviewers: petarj, jlpeyton, Hahnfeld, AndreyChurbanov
Reviewed By: petarj, AndreyChurbanov
Subscribers: AndreyChurbanov, sdardis, arichardson, atanasyan, jfb, jdoerfert, openmp-commits, llvm-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D60938
llvm-svn: 358911
Fix the test to run it really in SPMD mode without runtime. Previously
it was run in SPMD + full runtime mode and does not allow to cehck the
functionality correctly.
llvm-svn: 358902
Summary:
If the kernel is executed in SPMD mode and the L2+ parallel for region
with the dynamic scheduling is executed, dynamic scheduling functions
are called. They expect full runtime support, but SPMD kernels may be
executed without the full runtime. It leads to the runtime crash of the
compiled program. Patch fixes this problem + fixes handling of the
parallelism level in SPMD mode, which is required as part of this patch.
Reviewers: gtbercea, kkwli0, grokos
Subscribers: guansong, jdoerfert, openmp-commits, caomhin
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D60578
llvm-svn: 358442
This change replaces some of the assembly functions in z_Linux_asm.S
for inline asm in kmp.h. This allows better interaction with compiler
tools and sanitizers.
Differential Revision: https://reviews.llvm.org/D60423
llvm-svn: 358438
* Replace HBWMALLOC API with more general MEMKIND API, new functions
and variables added.
* Have libmemkind.so loaded when accessible.
* Redirect memspaces to default one except for high bandwidth which
is processed separately.
* Ignore some allocator traits e.g., sync_hint, access, pinned, while
others are processed normally e.g., alignment, pool_size, fallback,
fb_data, partition.
* Add tests for memory management
Patch by Andrey Churbanov
Differential Revision: https://reviews.llvm.org/D59783
llvm-svn: 357929
This patch cleans up the bookkeeping code for the load balancing dynamic mode.
When a thread is moved to or from the thread pool, the th_active_in_pool flag
and the __kmp_thread_pool_active_nth global counter are both updated. This
removes the need for the corrective code in the main wait loop. Another global
counter, __kmp_thread_pool_nth, was removed completely, as it was only used for
debugging, but was not under KMP_DEBUG.
Patch by Terry Wilmarth
Differential Revision: https://reviews.llvm.org/D59508
llvm-svn: 357927
Debug dump on large machine shows when many OpenMP threads (401 in total)
sleep on a barrier, one of the innermost nesting levels sleeps
on a child's b_arrived flag whose value is equal to 4 and is equal to
checker value. i.e., (1) sleep bit is 0, and (2) done_check() would
return true if called.
It is unclear how this might happen. It could be Windows Server 2016's
error of EnterCriticalSection / LeaveCriticalSection, or
error of WaitForSingleObject / SetEvent / ResetEvent, or
error in the library which is very difficult to find.
As a workaround, change INFINITE wait to timed wait, so that each
thread awakens each 5 seconds (the timeout was chosen arbitrary to not
disturb other threads much), check flag condition under the lock, and
either go to sleep again or stop sleeping as a result of the check.
Patch by Andrey Churbanov
Differential Revision: https://reviews.llvm.org/D59793
llvm-svn: 357722
The distribute clause needs an explicit push of a timer. The teams
clause needs a timer added and also, similarly to parallel, exchanged
with the serial timer when encountered so that serial regions are
counted properly.
Differential Revision: https://reviews.llvm.org/D59801
llvm-svn: 357621
On most platforms, certain compiler and linker flags have to be passed
when using pthreads, otherwise linking against libomp.so might fail with
undefined references to several pthread functions.
Use CMake's `find_package(Threads)` to determine these for standalone
builds, or take them (and optionally modify them) from the top-level
LLVM cmake files.
Also, On FreeBSD, ensure that libomp.so is linked against libm.so,
similar to NetBSD.
Adjust test cases with hardcoded `-lpthread` flag to use the common
build flags, which should now have the required pthread flags.
Reviewers: emaste, jlpeyton, krytarowski, mgorny, protze.joachim, Hahnfeld
Reviewed By: Hahnfeld
Subscribers: AndreyChurbanov, tra, EricWF, Hahnfeld, jfb, jdoerfert, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D59451
llvm-svn: 357618
At the moment, support for runtime debug output using the
OMPTARGET_DEBUG=1 environment variable is only available with
CMAKE_BUILD_TYPE=Debug builds. The patch allows setting it independently
using the LIBOMPTARGET_ENABLE_DEBUG option, which is enabled by default
depending on CMAKE_BUILD_TYPE. That is, unless this option is set
explicitly, nothing changes. This is the same mechanism used by LLVM for
LLVM_ENABLE_ASSERTIONS.
This patch also removes adding -g -O0 in debug builds, it should be
handled by cmake's CMAKE_{C|CXX}_FLAGS_DEBUG configuration option.
Idea by Hal Finkel
Differential Revision: https://reviews.llvm.org/D55952
llvm-svn: 356998
Summary:
While building the 8.0 releases on FreeBSD, I encountered the following
error in the regression tests, where ompt/misc/interoperability.cpp
failed to compile, with:
```
projects/openmp/runtime/test/ompt/misc/interoperability.cpp:7:10: fatal error: 'alloca.h' file not found
#include <alloca.h>
^~~~~~~~~~
```
Like on NetBSD, alloca(3) is defined in <stdlib.h> instead.
Reviewers: emaste, jlpeyton, krytarowski, mgorny, protze.joachim
Reviewed By: jlpeyton
Subscribers: jdoerfert, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D59736
llvm-svn: 356936
Summary:
[Split off from D59451 to get this fix in separately]
While building the 8.0 releases on FreeBSD, I encountered the following
warnings in openmp quite a few times:
```
In file included from projects/openmp/runtime/src/kmp_settings.cpp:27:
projects/openmp/runtime/src/kmp_wrapper_getpid.h:35:2: warning: #warning is a language extension [-Wpedantic]
#warning No gettid found, use getpid instead
^
projects/openmp/runtime/src/kmp_wrapper_getpid.h:35:2: warning: No gettid found, use getpid instead [-W#warnings]
2 warnings generated.
```
I added a gettid wrapper that uses FreeBSD's pthread_getthreadid_np(3)
function for this.
Reviewers: emaste, jlpeyton, krytarowski, mgorny, protze.joachim
Reviewed By: jlpeyton
Subscribers: jfb, jdoerfert, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D59735
llvm-svn: 356934
The GOMP sections interface uses schedule(dynamic) dispatch so it cannot
be assumed which thread executes the cancel and which thread executes
the cancellation point. This patch allows either thread to execute either
section.
llvm-svn: 356302
The following GCC intrinsics are not available on MIPS32:
__sync_fetch_and_add_8
__sync_fetch_and_and_8
__sync_fetch_and_or_8
__sync_val_compare_and_swap_8
Replace these with appropriate libatomic implementation.
Patch by Miodrag Dinic.
Differential Revision: https://reviews.llvm.org/D45691
llvm-svn: 355687
Summary:
The current install-clang-headers target installs clang's resource
directory headers. This is different from the install-llvm-headers
target, which installs LLVM's API headers. We want to introduce the
corresponding target to clang, and the natural name for that new target
would be install-clang-headers. Rename the existing target to
install-clang-resource-headers to free up the install-clang-headers name
for the new target, following the discussion on cfe-dev [1].
I didn't find any bots on zorg referencing install-clang-headers. I'll
send out another PSA to cfe-dev to accompany this rename.
[1] http://lists.llvm.org/pipermail/cfe-dev/2019-February/061365.html
Reviewers: beanz, phosek, tstellar, rnk, dim, serge-sans-paille
Subscribers: mgorny, javed.absar, jdoerfert, #sanitizers, openmp-commits, lldb-commits, cfe-commits, llvm-commits
Tags: #clang, #sanitizers, #lldb, #openmp, #llvm
Differential Revision: https://reviews.llvm.org/D58791
llvm-svn: 355340
Changing the default from -fPIC to -fno-PIC on PowerPC exposed an issue in
OpenMP for PowerPC.
The issue is reported here:
https://bugs.llvm.org/show_bug.cgi?id=40082
This is a fix for that issue.
Also removed the XFAIL from the two tests that were failing under -fno-PIC.
Differential Revision: https://reviews.llvm.org/D56286
llvm-svn: 355229
This change makes the runtime decide the intended use of each barrier
invocation, for the OMPT synchronization tool callbacks. The OpenMP 5.0
specification defines four possible barrier kinds -- implicit, explicit,
implementation, and just normal barrier.
Patch by Hansang Bae
Differential Revision: https://reviews.llvm.org/D58247
llvm-svn: 355140
Nest-var, OMP_NESTED, omp_set_nested()., and omp_get_nested() have been
deprecated in the 5.0 spec. Initial nesting info is now derived from
OMP_MAX_ACTIVE_LEVELS, OMP_NUM_THREADS, and OMP_PROC_BIND.
This patch deprecates the internal ICV that corresponds to nest-var, and
replaces it with the max-active-levels-var ICV to determine nesting. The
change still allows for use of OMP_NESTED (according to 5.0 changes),
omp_get_nested, and omp_set_nested, which have had deprecation messages
added to them. The change allows certain settings of OMP_NUM_THREADS,
OMP_PROC_BIND, and OMP_MAX_ACTIVE_LEVELS to turn on nesting, but
OMP_NESTED=0 will still force nesting to be off.
The runtime now prints informative messages about deprecation of
OMP_NESTED, omp_set_nested(), and omp_get_nested(), when those
environment variables or routines are used. It also prints deprecated
message in output for KMP_SETTINGS and OMP_DISPLAY_ENV for OMP_NESTED.
This patch also fixes OMP_DISPLAY_ENV output for OMP_TARGET_OFFLOAD.
Patch by Terry Wilmarth
Differential Revision: https://reviews.llvm.org/D58408
llvm-svn: 355138
This patch cleans up the yielding code and makes it optional. An
environment variable, KMP_USE_YIELD, was added. Yielding is still
on by default (KMP_USE_YIELD=1), but can be turned off completely
(KMP_USE_YIELD=0), or turned on only when oversubscription is detected
(KMP_USE_YIELD=2). Note that oversubscription cannot always be detected
by the runtime (for example, when the runtime is initialized and the
process forks, oversubscription cannot be detected currently over
multiple instances of the runtime).
Because yielding can be controlled by user now, the library mode
settings (from KMP_LIBRARY) for throughput and turnaround have been
adjusted by altering blocktime, unless that was also explicitly set.
In the original code, there were a number of places where a double yield
might have been done under oversubscription. This version checks
oversubscription and if that's not going to yield, then it does
the spin check.
Patch by Terry Wilmarth
Differential Revision: https://reviews.llvm.org/D58148
llvm-svn: 355120
Summary:
This patch adds a more sophisticated team reduction scheme to the OpenMP libomptarget-nvptx runtime.
The scheme uses a fixed size global memory buffer whose length can be adjusted via compiler flag:
```
-fopenmp-cuda-teams-reduction-recs-num=1024
```
The global buffer is a structure of arrays (with default size of 1024 each and controlled by the above flag), one array for each reduction variable.
Values in the buffer are processed by the last team to finish executing the body of the target region.
In addition to adding support for the new flag, the compiler also emits special functions used for the reduction of the intermediate reduction values. These changes will be added in a separate compiler patch following this one.
Reviewers: ABataev, caomhin
Reviewed By: ABataev
Subscribers: guansong, jfb, jdoerfert, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D58409
llvm-svn: 354471
This patch adds the new 5.0 API function omp_get_supported_active_levels().
Patch by Terry Wilmarth
Differential Revision: https://reviews.llvm.org/D58211
llvm-svn: 354368
Remove fatal error messages from the cancellation API for GOMP
Add __kmp_barrier_gomp_cancel() to implement cancellation of parallel regions.
This new function uses the linear barrier algorithm with a cancellable
nonsleepable wait loop.
Differential Revision: https://reviews.llvm.org/D57969
llvm-svn: 354367
The thread-limit-var and omp_get_thread_limit API was not perfectly handled for
teams construct. Now, when modified by thread_limit clause, omp_get_thread_limit
reports the correct value. In addition, the value is restored when leaving the
teams construct to what it was in the encountering context.
This is done partly by creating the notion of a Contention Group root (CG root)
that keeps track of the thread at the root of each separate CG, the
thread-limit-var associated with the CG, and associated counter of active
threads within the contention group.
thread-limits are passed from master to worker threads via an entry in the ICV
data structure. When a "contention group switch" occurs, a new CG root record is
made and passed from master to worker. A thread could potentially have several
CG root records if it encounters multiple nested teams constructs (but at the
moment the spec doesn't allow for nested teams, so the most one could have
currently is 2). The master of the teams masters gets the thread-limit clause
value stored to its local ICV structure, and the other teams masters copy it
from the master. The thread-limit is set from that ICV copy and restored to the
ICV copy when entering and leaving the teams construct.
This change also fixes a bug when the top-level teams construct team gets
reused, and OMP_DYNAMIC was true, which can cause the expected size of this team
to be smaller than what was actually allocated. The fix updates the size of the
team after its threads were reserved.
Patch by Terry Wilmarth
Differential Revision: https://reviews.llvm.org/D56804
llvm-svn: 353747
%s refers to the test file in the source tree. This was accidentally added in
r351197 / 2b46d30 ("[OMPT] Second chunk of final OMPT 5.0 interface updates").
Differential Revision: https://reviews.llvm.org/D58002
llvm-svn: 353715
Summary:
As @david2050 commented, changes introduced by https://reviews.llvm.org/D56397 break builds for older compilers
which don't support `__has(_cpp)_attribute`. This is a fix for the break.
Reviewers: protze.joachim, jlpeyton, AndreyChurbanov, Hahnfeld, david2050
Subscribers: openmp-commits, david2050
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D57851
llvm-svn: 353538
The three switch fallthrough generate a warning with -Wimplicit-fallthrough.
Two are documented as fallthrough, one is not, but I think the intention is to also fallthrough in kmp_tasking.cpp.
Not sure whether kmp.h is the best place to define the macro.
Reviewers: jlpeyton, AndreyChurbanov, Hahnfeld
Reviewed By: jlpeyton
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D56397
llvm-svn: 353052
Redo after revert by hans. The wrong include in one test is fixed.
Make sure that OMPT is enabled in runtime entry points that access internals
of the runtime. Else, return an appropiate value indicating an error or that
the data is not available.
Patch provided by @sconvent
Reviewers: jlpeyton, omalyshe, hbae, Hahnfeld, joachim.protze
Reviewed By: joachim.protze
Tags: #openmp, #ompt
Differential Revision: https://reviews.llvm.org/D47717
llvm-svn: 352611
This fixes most references to the paths:
llvm.org/svn/
llvm.org/git/
llvm.org/viewvc/
github.com/llvm-mirror/
github.com/llvm-project/
reviews.llvm.org/diffusion/
to instead point to https://github.com/llvm/llvm-project.
This is *not* a trivial substitution, because additionally, all the
checkout instructions had to be migrated to instruct users on how to
use the monorepo layout, setting LLVM_ENABLE_PROJECTS instead of
checking out various projects into various subdirectories.
I've attempted to not change any scripts here, only documentation. The
scripts will have to be addressed separately.
Additionally, I've deleted one document which appeared to be outdated
and unneeded:
lldb/docs/building-with-debug-llvm.txt
Differential Revision: https://reviews.llvm.org/D57330
llvm-svn: 352514
As the codebase is now under the Apache 2.0 license with LLVM
Exceptions, and all Arm's contributions, past or future, are under that
new license, this Arm specific words in LICENSE.txt are no longer
needed.
llvm-svn: 352377
all missed!
Thanks to Alex Bradbury for pointing this out, and the fact that I never
added the intended `legacy` anchor to the developer policy. Add that
anchor too. With hope, this will cause the links to all resolve
successfully.
llvm-svn: 351731
to reflect the new license. These used slightly different spellings that
defeated my regular expressions.
We understand that people may be surprised that we're moving the header
entirely to discuss the new license. We checked this carefully with the
Foundation's lawyer and we believe this is the correct approach.
Essentially, all code in the project is now made available by the LLVM
project under our new license, so you will see that the license headers
include that license only. Some of our contributors have contributed
code under our old license, and accordingly, we have retained a copy of
our old license notice in the top-level files in each project and
repository.
llvm-svn: 351648
This installs the new developer policy and moves all of the license
files across all LLVM projects in the monorepo to the new license
structure. The remaining projects will be moved independently.
Note that I've left odd formatting and other idiosyncracies of the
legacy license structure text alone to make the diff easier to read.
Critically, note that we do not in any case *remove* the old license
notice or terms, as that remains necessary until we finish the
relicensing process.
I've updated a few license files that refer to the LLVM license to
instead simply refer generically to whatever license the LLVM project is
under, basically trying to minimize confusion.
This is really the culmination of so many people. Chris led the
community discussions, drafted the policy update and organized the
multi-year string of meeting between lawyers across the community to
figure out the strategy. Numerous lawyers at companies in the community
spent their time figuring out initial answers, and then the Foundation's
lawyer Heather Meeker has done *so* much to help refine and get us ready
here. I could keep going on, but I just want to make sure everyone
realizes what a huge community effort this has been from the begining.
Differential Revision: https://reviews.llvm.org/D56897
llvm-svn: 351631
and also the follow-up r351315.
The new test is failing on the buildbots.
> Make sure that OMPT is enabled in runtime entry points that access internals
> of the runtime. Else, return an appropiate value indicating an error or that
> the data is not available.
>
> Patch provided by @sconvent
>
> Reviewers: jlpeyton, omalyshe, hbae, Hahnfeld, joachim.protze
>
> Reviewed By: joachim.protze
>
> Tags: #openmp, #ompt
>
> Differential Revision: https://reviews.llvm.org/D47717
llvm-svn: 351431
Add omp_pause_resource and omp_pause_resource_all API and enum, plus stub for
internal implementation. Implemented callable helper function to do local pause,
and added basic functionality for hard and soft pause.
Patch by Terry Wilmarth
Differential Revision: https://reviews.llvm.org/D55078
llvm-svn: 351372
The compiler warns about an unused variable/statement:
runtime/src/kmp_affinity.cpp:4958:18: warning: statement has no effect [-Wunused-value]
KA_TRACE(1000, ; {
^
runtime/src/kmp_debug.h:84:24: note: in definition of macro 'KA_TRACE'
__kmp_debug_printf x; \
^
Instead of the unused reference to this function, this patch now calls the function
with an empty string. The call to this function should have no effect.
Patch provided by joachim.protze
Reviewers: jlpeyton, hbae, AndreyChurbanov
Reviewed By: AndreyChurbanov
Tags: #openmp, #ompt
Differential Revision: https://reviews.llvm.org/D56775
llvm-svn: 351323
Make sure that OMPT is enabled in runtime entry points that access internals
of the runtime. Else, return an appropiate value indicating an error or that
the data is not available.
Patch provided by @sconvent
Reviewers: jlpeyton, omalyshe, hbae, Hahnfeld, joachim.protze
Reviewed By: joachim.protze
Tags: #openmp, #ompt
Differential Revision: https://reviews.llvm.org/D47717
llvm-svn: 351311
Using proc_bind clause on a nested #pragma omp parallel region
with KMP_AFFINITY set causes an assertion error. This assertion occurs because
the place-partition-var is not properly initialized in the nested master threads.
Trying to get an intuitive result with KMP_AFFINITY + proc_bind is difficult
because of how the KMP_AFFINITY gtid-to-place mapping occurs. This
patch creates an initial place list no matter what affinity mechanism is used.
For KMP_AFFINITY, the place-partition-var is initialized to all the places.
Differential Revision: https://reviews.llvm.org/D55795
llvm-svn: 351227
This change fixes the sanity issue reported in Bug 40042.
Lock function definitions for the three lock kinds were added
to disambiguate calls to the lock functions done directly and indirectly.
Bugzilla: https://bugs.llvm.org/show_bug.cgi?id=40042
Patch by Hansang Bae
Differential Revision: https://reviews.llvm.org/D56103
llvm-svn: 351224
Fix ninja build detect test compiler failed under windows.
Patch by Peiyuan Song
Differential Revision: https://reviews.llvm.org/D53479
llvm-svn: 351223
Make __ompt_implicit_task_end a static function and remove the inline part. Remove
pId variable that is unused. This fixes small regression in SPEC kdtree benchmark.
Also reformat some of __ompt_implicit_task_end.
Differential Revision: https://reviews.llvm.org/D55788
llvm-svn: 351221
The omp-tools.h file is generated from the OpenMP spec to ensure that the interface
is implemented as specified.
The other changes are necessary to update the interface implementation to the
final version as published in 5.0.
The omp-tools.h header was previously called ompt.h, currently a copy under this name
is installed for legacy tools.
Patch partially perpared by @sconvent
Reviewers: AndreyChurbanov, hbae, Hahnfeld
Reviewed By: hbae
Tags: #openmp, #ompt
Differential Revision: https://reviews.llvm.org/D55579
llvm-svn: 351197
Summary:
Two things:
1. Those two variables had the wrong sigdness, which was resulting in "sign mismatch in comparison" warning.
2. The whole `kmp_debugger.cpp` wasn't being built, or rather, it was being built as-if `USE_DEBUGGER` was off,
thus, nothing provided the definition of `__kmp_omp_debug_struct_info`, `__kmp_debugging`.
Makes sense, because `USE_DEBUGGER` is set in `kmp_config.h`, which is not included explicitly.
It is included by `kmp.h`, but that one is only included inside of the `#if USE_DEBUGGER` block..
I *think* this is the only source file with this issue,
everything else seem to `#include` either `kmp.h` or `kmp_config.h`.
The alternative solution would be to add `add_compile_options(-include kmp_config.h)` in CMake.
I did verify that `__kmp_omp_debug_struct_info` becomes available with this patch.
Fixes [[ https://bugs.llvm.org/show_bug.cgi?id=38612 | PR38612 ]].
Reviewers: AndreyChurbanov, jlpeyton, Hahnfeld
Reviewed By: jlpeyton
Subscribers: guansong, jfb, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D55783
llvm-svn: 351019
Summary: Replace existing infrastructure for tracking parallel level using global memory with a per-team shared memory variable. This minimizes the impact of the overhead of tracking the parallel level for non-nested cases.
Reviewers: ABataev, caomhin
Reviewed By: ABataev
Subscribers: guansong, openmp-commits
Differential Revision: https://reviews.llvm.org/D55773
llvm-svn: 350747
Summary:
Previous implementation may cause the runtime crash when the number of
teams is > 1024. Patch fixes this problem + reduces number of the atomic
operations by 32 times.
Reviewers: grokos, gtbercea, kkwli0
Subscribers: guansong, jfb, openmp-commits, caomhin
Differential Revision: https://reviews.llvm.org/D56332
llvm-svn: 350524
Summary:
Reduced number of the used register + improved performance propagating
the information about current execution/data sharing mode directly from
the compiler, where it is possible.
In some cases, it requires new/reworked interfaces of the runtime
external functions. Old functions are marked as deprecated.
Reviewers: grokos, gtbercea, kkwli0
Subscribers: guansong, jfb, openmp-commits, caomhin
Differential Revision: https://reviews.llvm.org/D56278
llvm-svn: 350405
The OpenMP runtime's cmake scripts do not correctly locate the
libdevice that the Debian/Ubuntu package nvidia-cuda-toolkit currently
includes, at least on my Ubuntu 18.04.1 installation. This patch
fixes that for me.
This problem was discussed at length in D55269. D40453 added a
similar adjustment in clang, but reviewers of D55269 concluded that,
for the OpenMP runtime, the right place to address this problem is in
cmake's CUDA support. However, it was also suggested we could add a
workaround to OpenMP's cmake scripts now. This patch contains such a
workaround, which I've tried to design so that it will have no harmful
effect if cmake improves in the future.
nvidia-cuda-toolkit also needs improvements because its intended
monolithic CUDA tree shim, /usr/lib/cuda, has many empty directories,
such as bin. I reported that at:
<https://bugs.launchpad.net/ubuntu/+source/nvidia-cuda-toolkit/+bug/1808999>
Reviewed By: grokos
Differential Revision: https://reviews.llvm.org/D55588
llvm-svn: 350377
Add omp_get_device_num() function for 5.0 which returns the number of the
device the current thread is running on. Currently, we are leaving it to the
compiler to handle this properly if it is called inside target.
Also, did some cleanup and updating of duplicate device API functions (in both
libomp and libomptarget) to make them into weak functions that check for the
symbol from libomptarget, and will call the version in libomptarget if it is
present. If any additional device API functions are implemented also in
libomptarget in the future, we should add the dlsym calls to the host functions.
Also, if the omp_target_* functions are to be implemented for the host (this has
been requested), they should attempt to call the libomptarget versions as well.
Patch by Terry Wilmarth
Differential Revision: https://reviews.llvm.org/D55578
llvm-svn: 350352
Summary:
One of the LLVM optimizations, split critical edges, also clones tail
instructions. This is a dangerous operation for __syncthreads()
functions and this transformation leads to undefined behavior or
incorrect results. Patch fixes this problem by replacing __syncthreads()
function with the assembler instruction, which cost is too high and
wich cannot be copied.
Reviewers: grokos, gtbercea, kkwli0
Subscribers: guansong, openmp-commits, caomhin
Differential Revision: https://reviews.llvm.org/D56274
llvm-svn: 350333
Summary:
Avoid using of the atomic loop to wait for the completion of the
data-sharing interface initialization, use __shfl_sync instead for the
communication within the warp to signal other threads in the warp about
completion of the initialization.
Reviewers: gtbercea, kkwli0, grokos
Subscribers: guansong, jfb, caomhin, openmp-commits
Differential Revision: https://reviews.llvm.org/D56100
llvm-svn: 350129
Summary:
At high optimization level asserts lead to some unexpected results
because of auto-inserted unreachable instructions. This outlining
prevents some of such dangerous optimizations and leads to better
stability.
Reviewers: gtbercea, kkwli0, grokos
Subscribers: guansong, caomhin, openmp-commits
Differential Revision: https://reviews.llvm.org/D56101
llvm-svn: 350128
Fix the newly-added tests to use %python substitution in order to use
the correct path to Python interpreter. Otherwise, they fail on NetBSD
where there is no 'python', just 'pythonX.Y'.
Differential Revision: https://reviews.llvm.org/D56048
llvm-svn: 350001
XFAIL two tests that fail on PowerPC LE Linux due
to the change of default from PIC to no-PIC on that
platform.
A Bug has been opened for this:
https://bugs.llvm.org/show_bug.cgi?id=40082
The tests are:
runtime/test/ompt/misc/control_tool.c
runtime/test/ompt/synchronization/taskwait.c
llvm-svn: 349512
This patch updates the implementation of the ompt_frame_t, ompt_wait_id_t
and ompt_state_t. The final version of the OpenMP 5.0 spec added the "t"
for these types.
Furthermore the structure for ompt_frame_t changed and allows to specify
that the reenter frame belongs to the runtime.
Patch partially prepared by Simon Convent
Reviewers: hbae
llvm-svn: 349458
Summary:
I have discovered this because i wanted to experiment with
building static libomp (with openmp-4.0 support only)
for debugging purposes.
There are three kinds of problems here:
1. `__kmp_compare_and_store_acq()` simply does not exist.
It was added in D47903 by @jlpeyton.
I'm guessing `__kmp_atomic_compare_store_acq()` was meant.
2. In `__kmp_is_ticket_lock_initialized()`,
`lck->lk.initialized` is `std::atomic<bool>`,
while `lck` is `kmp_ticket_lock_t *`.
Naturally, they can't be equality-compared.
Either, it should return the value read from `lck->lk.initialized`,
or do what `__kmp_is_queuing_lock_initialized()` does,
compare the passed pointer with the field in the struct
pointed by the pointer. I think the latter is correct-er choice here.
3. Tests were not versioned.
They assume that `LIBOMP_OMP_VERSION` is at the latest version.
This does not touch LIBOMP_OMP_VERSION=30. That is still broken.
Reviewers: jlpeyton, Hahnfeld, AndreyChurbanov
Reviewed By: AndreyChurbanov
Subscribers: guansong, jfb, openmp-commits, jlpeyton
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D55496
llvm-svn: 349260
The value returned by __kmp_now_nsec() can overflow 32-bit values causing
incorrect values to be returned. The overflow can end up causing a divide
by zero error because in __kmp_initialize_system_tick(), the value
(__kmp_now_nsec() - nsec) can end up being much larger than the numerator:
1e6 * (delay + (now - goal))
during a pathological timing where the current time calculated is much larger
than nsec. When this happens, the value of __kmp_ticks_per_msec is set to zero
which is then used as the denominator in the KMP_NOW_MSEC() macro leading to
the divide by zero error.
Differential Revision: https://reviews.llvm.org/D55300
llvm-svn: 349090
This patch adds the affinity format functionality introduced in OpenMP 5.0.
This patch adds: Two new environment variables:
OMP_DISPLAY_AFFINITY=TRUE|FALSE
OMP_AFFINITY_FORMAT=<string>
and Four new API:
1) omp_set_affinity_format()
2) omp_get_affinity_format()
3) omp_display_affinity()
4) omp_capture_affinity()
The affinity format functionality has two ICV's associated with it:
affinity-display-var (bool) and affinity-format-var (string).
The affinity-display-var enables/disables the functionality through the
envirable OMP_DISPLAY_AFFINITY. The affinity-format-var is a formatted
string with the special field types beginning with a '%' character
similar to printf
For example, the affinity-format-var could be:
"OMP: host:%H pid:%P OStid:%i num_threads:%N thread_num:%n affinity:{%A}"
The affinity-format-var is displayed by every thread implicitly at the beginning
of a parallel region when any thread's affinity has changed (including a brand
new thread being spawned), or explicitly using the omp_display_affinity() API.
The omp_capture_affinity() function can capture the affinity-format-var in a
char buffer. And omp_set|get_affinity_format() allow the user to set|get the
affinity-format-var explicitly at runtime. omp_capture_affinity() and
omp_get_affinity_format() both return the number of characters needed to hold
the entire string it tried to make (not including NULL character). If not
enough buffer space is available,
both these functions truncate their output.
Differential Revision: https://reviews.llvm.org/D55148
llvm-svn: 349089
Disable KMP_HAVE_QUAD when building via gcc on NetBSD system,
as the build fails due to unimplemented builtins:
.../kmp_atomic.cpp.o: In function `__kmpc_atomic_cmplx16_mul':
.../kmp_atomic.cpp:1332: undefined reference to `__multc3'
.../kmp_atomic.cpp.o: In function `__kmpc_atomic_cmplx16_div':
.../kmp_atomic.cpp:1334: undefined reference to `__divtc3'
...
Differential Revision: https://reviews.llvm.org/D55478
llvm-svn: 348886
Switch NetBSD from reading /proc (which is broken) to getloadavg()
(which is already used by Darwin). NetBSD discourages using procfs
in favor of system API calls.
Differential Revision: https://reviews.llvm.org/D55486
llvm-svn: 348885
Summary:
Use the sysctl(3) function to check whether an address is mapped
into the address space.
Reviewers: mgorny, joerg, #openmp
Reviewed By: mgorny
Subscribers: openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D55549
llvm-svn: 348874
Summary: _lwp_self() returns current Thread Id in a numeric version on NetBSD.
Reviewers: joerg, mgorny, #openmp
Reviewed By: mgorny
Subscribers: llvm-commits, openmp-commits, #openmp
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D55497
llvm-svn: 348873
Increase the range for omp_get_wtick() test to allow for 0.01
(from <0.01). This is needed for NetBSD where it returns exactly that
value due to CLOCKS_PER_SEC being 100. This should not cause
a significant difference from e.g. FreeBSD where it is 128,
and especially from Linux where CLOCKS_PER_SEC is apparently meaningless
and sysconf(_SC_CLK_TCK) gives 100 as well.
Differential Revision: https://reviews.llvm.org/D55493
llvm-svn: 348857
On NetBSD, alloca() is in stdlib.h and there is no alloca.h. Adjust
the includes appopriately.
Differential Revision: https://reviews.llvm.org/D55487
llvm-svn: 348856
Pass `-n -s` instead of `--numeric --stable` to sort(1), as long options
are not supported by NetBSD sort implementation. `-n` is defined
by POSIX, so it should be fully portable. `-s` is used consistently
at least in GNU sort and FreeBSD sort, and I honestly doubt it would
cause issues with any other implementation supporting `--stable`.
Differential Revision: https://reviews.llvm.org/D55479
llvm-svn: 348855
Prefer using '-std=gnu++11' over '-std=c++11' when available, as NetBSD
exposes the correct alloca() implementation only with gnu* C/C++
standards.
Differential Revision: https://reviews.llvm.org/D55477
llvm-svn: 348854
Fix two build issues:
1) Recent commit 348756 accidentally included Unix clang compilers
to use immintrin.h when only clang-cl should be using it leading
to the following error:
openmp-llvm/runtime/src/kmp_lock.cpp:2035:25: error: always_
inline function '_xbegin' requires target feature 'rtm', but would be inlined into function
'__kmp_test_adaptive_lock_only' that is compiled without support for 'rtm'
kmp_uint32 status = _xbegin();
This patch changes the guard to use immintrin.h to only use clang-cl instead of all clang
2) gcc-8 gives a warning about multiline comment in kmp_runtime.cpp:
This patch just changes it to a two line comment
openmp-llvm/runtime/src/kmp_runtime.cpp:7697:8: warning: multi-line comment [-Wcomment]
#endif // KMP_OS_LINUX || KMP_OS_DRAGONFLY || KMP_OS_FREEBSD || KMP_OS_NETBSD \
llvm-svn: 348783
Summary:
Use the original shuffle implementation for __kmpc_shuffle_int64 since
default implementation uses the same implementation.
Reviewers: gtbercea
Subscribers: guansong, caomhin, openmp-commits
Differential Revision: https://reviews.llvm.org/D55514
llvm-svn: 348772
Summary:
Shuffle on 64bit data is allowed only for CUDA >= 9.0. Also, fixed the
constant for the mask, need one extra L in the end.
Reviewers: gtbercea, kkwli0
Subscribers: guansong, caomhin, openmp-commits
Differential Revision: https://reviews.llvm.org/D55440
llvm-svn: 348758
Summary: This patch permits OpenMP to build and work (with both gcc and clang) on OpenBSD. It mostly follows what was done for FreeBSD and NetBSD, except OpenBSD does not have pthread_getattr_np support, so it follows OS X in that one instance.
Reviewers: #openmp, krytarowski
Reviewed By: krytarowski
Subscribers: guansong, jfb, emaste, mgorny, krytarowski, #openmp
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D34280
llvm-svn: 348726
Summary:
Additions mostly follow FreeBSD and NetBSD and are not intrusive.
There is similar patch for OpenBSD: https://reviews.llvm.org/D34280
The -lm was being omitted due to -Wl,--as-needed in cmake rule, similar patch is in freebsd-ports/devel/llvm-devel port.
Simple OpenMP programs compile and work as expected:
$ clang-devel ~/omp_hello.c -fopenmp -I/usr/local/llvm-devel/include
$ LD_LIBRARY_PATH=/usr/local/llvm-devel/lib OMP_NUM_THREADS=100 ./a.out
The assertion in LLVMgold.so when -fopenmp was used together with -flto in 20170524 snapshot is no longer triggered on current svn-trunk and works fine as in llvm-4.0 with our local patches.
Reviewers: #openmp, krytarowski
Reviewed By: krytarowski
Subscribers: dexonsmith, jfb, krytarowski, guansong, gregrodgers, emaste, mgorny, mehdi_amini
Differential Revision: https://reviews.llvm.org/D35129
llvm-svn: 348725
Summary:
Introduced special noinline function log that allows to save some
registers for optimized builds but with enabled logging. Also, it
increases the stability of the optimized builds with inlined runtime.
Reviewers: gtbercea, kkwli0
Reviewed By: gtbercea
Subscribers: caomhin, guansong, openmp-commits
Differential Revision: https://reviews.llvm.org/D55436
llvm-svn: 348606
Summary:
According to the standard, after memory flushing the changes in the
memory must be visible to all the threads in all teams. Patch fixes
this.
Reviewers: gtbercea, kkwli0
Subscribers: guansong, jfb, caomhin, openmp-commits
Differential Revision: https://reviews.llvm.org/D55370
llvm-svn: 348491
Summary:
Reworked runtime to make it compatible with the requirements of the
original runtime library. Also, simplified some code to reduce number of
function calls.
Reviewers: gtbercea, kkwli0
Subscribers: guansong, jfb, caomhin, openmp-commits
Differential Revision: https://reviews.llvm.org/D55130
llvm-svn: 348003
There is a conflict between libomptarget and libomp concerning some of the
standard OpenMP device API which needs further intestigation.
llvm-svn: 347932
This patch adds __kmpc_omp_reg_task_with_affinity to register affinity
information for tasks. For now, the affinity information is not used,
and the function always succeeds. This also adds the kmp_task_affinity_info_t
structure to store the task affinity information.
Patch by Terry Wilmarth
Differential Revision: https://reviews.llvm.org/D55026
llvm-svn: 347907
This change renames ompt_mutex_impl_unknown to ompt_mutex_impl_none,
following the name change in the specification.
Patch by Hansang Bae
Differential Revision: https://reviews.llvm.org/D54347
llvm-svn: 347802
* Fix calculation of string length.
* Remove NULL-check of pointer which has been dereferenced.
Patch by Andrey Churbanov
Differential Revision: https://reviews.llvm.org/D54948
llvm-svn: 347801
There is low probability that array th_hot_teams can be
accessed out of bound (when many nested levels are requested
to keep hot teams via KMP_HOT_TEAMS_MAX_LEVEL). The patch
adds the check of index that fixes the problem.
Patch by Andrey Churbanov
Differential Revision: https://reviews.llvm.org/D54950
llvm-svn: 347800
Add omp_get_device_num() function for 5.0 which returns the number of the device
the current thread is running on. Also, did some cleanup and updating of device
API functions to make them into weak functions that should be replaced with
libomptarget functions when libomptarget is present.
Patch by Terry Wilmarth
Differential Revision: https://reviews.llvm.org/D54342
llvm-svn: 347799
Summary: To enable the compiler to optimize parts of the function that are not needed when runtime can be omitted, a new version of the SPMD deinit kernel function is needed. This function takes the runtime required flag as an argument.
Reviewers: ABataev, kkwli0, caomhin
Reviewed By: ABataev
Subscribers: guansong, openmp-commits
Differential Revision: https://reviews.llvm.org/D54969
llvm-svn: 347714
Summary:
Added functions __kmpc_nvptx_teams_reduce_nowait_simple and
__kmpc_nvptx_teams_end_reduce_nowait_simple to implement basic support
for reductions across the teams.
Reviewers: gtbercea, kkwli0
Subscribers: guansong, jfb, caomhin, openmp-commits
Differential Revision: https://reviews.llvm.org/D54967
llvm-svn: 347710
Summary: Refactor the checking for SPMD mode and whether the runtime is initialized or not. This uses constant flags which enables the runtime to optimize out unused sections of code that depend on these flags.
Reviewers: ABataev, caomhin
Reviewed By: ABataev
Subscribers: guansong, jfb, openmp-commits
Differential Revision: https://reviews.llvm.org/D54960
llvm-svn: 347698
Summary:
The base pointer for the lambda mapping must point to the lambda capture
placement and pointer must point to the captured variable itself. Patch
fixes this problem.
Reviewers: gtbercea
Subscribers: guansong, openmp-commits, kkwli0, caomhin
Differential Revision: https://reviews.llvm.org/D54260
llvm-svn: 346407
Summary:
The previously used combination `PTR_AND_OBJ | PRIVATE` could be used
for mapping of some data in Fortran. Changed it to `PTR_AND_OBJ |
LITERAL`.
Reviewers: gtbercea
Subscribers: guansong, caomhin, openmp-commits
Differential Revision: https://reviews.llvm.org/D54035
llvm-svn: 345981
Summary:
Current globalization scheme works correctly only for SPMD+lightweight
runtime mode and does not work for full runtime. Patch improves support
for the globalization scheme + reduces global memory consumption in
lightweight runtime mode.
Patch adds runtime functions to work with the statically allocated
global memory. It allows to improve performance and memory consumption.
This global memory must be allocated by the compiler.
Reviewers: grokos, kkwli0, gtbercea, caomhin
Subscribers: guansong, jfb, openmp-commits
Differential Revision: https://reviews.llvm.org/D53943
llvm-svn: 345976
Summary: In the case of coalesced global records, we need to push the exact data size passed in. This patch fixes this by outlining the common functionality of the previous push function and by adding a separate entry point for coalesced pushes. The pop function remains unchanged.
Reviewers: ABataev, grokos, caomhin
Reviewed By: ABataev, grokos
Subscribers: jholewinski, cfe-commits, Hahnfeld, guansong, jfb, openmp-commits
Differential Revision: https://reviews.llvm.org/D53141
llvm-svn: 345867
Summary:
Added support for correct mapping of variables captured by reference in
lambdas. That kind of mapping may appear only in target-executable
regions and must follow the original lambda or another lambda capture
for the same lambda.
The expected data: base address - the address of the lambda, begin
pointer - pointer to the address of the lambda capture, size - size of
the captured variable.
When OMP_TGT_MAPTYPE_PTR_AND_OBJ mapping type is seen in
target-executable region, the target address of the last processed item
is taken as the address of the original lambda `tgt_lambda_ptr`. Then,
the pointer to capture on the device is calculated like `tgt_lambda_ptr
+ (host_begin_pointer - host_begin_base)` and the target-based address
of the original variable (which host address is
`*(void**)begin_pointer`) is written to that pointer.
Reviewers: kkwli0, gtbercea, grokos
Subscribers: openmp-commits
Differential Revision: https://reviews.llvm.org/D51107
llvm-svn: 345608
Initializing an ompt_data_t object using the pointer union member is potentially
unsafe in 32-bit programs. This change fixes the issue
by using the constant, ompt_data_none.
Patch by Hansang Bae
Differential Revision: https://reviews.llvm.org/D52046
llvm-svn: 343785
On Windows, child workers are terminated by the parent during the normal
program exit process (ExitProcess()) and they are not able to finish generating
their OpenMP events. We can force manual library shut down in __kmpc_end() to
fix this at least for the cases where __kmpc_end() is properly inserted.
Patch by Hansang Bae
Differential Revision: https://reviews.llvm.org/D52628
llvm-svn: 343619
If the user requested LIBOMPTARGET_NVPTX_DEBUG, include asserts in
the bitcode library. Everything else will have very unpleasent
effects because asserts will appear when falling back to the static
library libomptarget-nvptx.a.
Differential Revision: https://reviews.llvm.org/D52701
llvm-svn: 343477
Pass in the correct value of isRuntimeUninitialized() which solves
parallel reductions as reported on the mailing list.
For reference: r333285 did the same for loop scheduling.
Differential Revision: https://reviews.llvm.org/D52725
llvm-svn: 343476
Patch suggested by Kelvin Li: removed optional "kind=" part of kind-selector
for variables with long names and kind names.
Differential Revision: https://reviews.llvm.org/D52712
llvm-svn: 343475
NVPTX requires addresses of pointer locations to be 8-byte aligned
or there will be an exception during runtime.
This could happen without this patch as shown in the added test:
getId() requires 4 byte of stack and putValueInParallel() uses 16
bytes to store the addresses of the captured variables.
Differential Revision: https://reviews.llvm.org/D52655
llvm-svn: 343402
According to OpenMP 4.5, p250:12-14:
If the requested nest level is outside the range of 0 and the
nest level of the current thread, as returned by the omp_get_level
routine, the routine returns -1.
The SPMD code path will need a similar fix.
Differential Revision: https://reviews.llvm.org/D51787
llvm-svn: 343401
Clang trunk will serialize nested parallel regions. Check that this
is correctly reflected in various API methods.
Differential Revision: https://reviews.llvm.org/D51786
llvm-svn: 343382
There is no support and according to the OpenMP 4.5, p238:7-9:
For implementations that do not support dynamic adjustment
of the number of threads this routine has no effect: the
value of dyn-var remains false.
Add a test that cancellation and nested parallelism aren't
supported either.
Differential Revision: https://reviews.llvm.org/D51785
llvm-svn: 343381
If there is no num_threads() clause we must consider the
nthreads-var ICV. Its value is set by omp_set_num_threads()
and can be queried using omp_get_max_num_threads().
The rewritten code now closely resembles the algorithm given
in the OpenMP standard.
Differential Revision: https://reviews.llvm.org/D51783
llvm-svn: 343380
infinite loop on removing non-mapped pointer-with-object.
Added test to check that libomptarget does not cause infinite loop when
trying to unmap the pointer-with-object data that was not previously
mapped.
llvm-svn: 343344
This patch also introduces testing for libomptarget-nvptx
which has been missing until now. I propose to add tests for
all bugs that are fixed in the future.
The target check-libomptarget-nvptx is not run by default because
- we can't determine if there is a GPU plugged into the system.
- it will require the latest Clang compiler. Keeping compatibility
with older releases would prevent testing newer code generation
developed in trunk.
Differential Revision: https://reviews.llvm.org/D51687
llvm-svn: 343324
This patch puts the __kmpc_critical_with_hint function in dllexports
and also replaces some OMP_45_ENABLED to OMP_50_ENABLED
Differential Revision: https://reviews.llvm.org/D52380
llvm-svn: 343143
Balanced affinity only updated the thread's affinity with the operating system.
This change also has the thread's private mask reflect that change as well so
that any API that probes the thread's affinity mask will report the correct
mask value.
Differential Revision: https://reviews.llvm.org/D52379
llvm-svn: 343142
This patch updates the ittnotify sources to the latest
corresponding with Intel(R) VTune(TM) Amplifier 2018
Differential Revision: https://reviews.llvm.org/D52378
llvm-svn: 343139
This change improves the performance of 376.kdtree by giving the compiler an
opportunity to do inlining and other optimizations for the call path,
__kmpc_omp_task_complete_if0()->__kmp_task_finish(), which is one of the hot
paths in the program; some functions in kmp_taskdeps.cpp were moved to the new
header file, kmp_taskdeps.h to achieve this.
Patch by Hansang Bae
Differential Revision: https://reviews.llvm.org/D51889
llvm-svn: 343138
This change includes miscellaneous improvements as follows:
1) Added ompt_get_proc_id() implementation for Windows
2) Added parser and print tool for omp-tool-var, just in case it needs
to be printed (OMP_DISPLAY_ENV)
3) omp_control_tool is exported on Windows
Patch by Hansang Bae
Differential Revision: https://reviews.llvm.org/D50538
llvm-svn: 343137
Summary: NFC - just fixing a bug: the empty slot test was before the re-setting of the Stack pointer.
Reviewers: ABataev, caomhin, Hahnfeld
Reviewed By: ABataev
Subscribers: guansong, openmp-commits
Differential Revision: https://reviews.llvm.org/D52122
llvm-svn: 343006
Summary:
There is currently no supported situation where the warp master is not the first thread in the warp.
This also avoids the device execution from hanging on Volta GPUs when ballot_sync is called by a number of threads that is less that the size of a warp.
Reviewers: ABataev, caomhin, grokos
Reviewed By: grokos
Subscribers: guansong, openmp-commits
Differential Revision: https://reviews.llvm.org/D50188
llvm-svn: 342972
Summary:
We need the support for per-team shared variables to support codegen for
lastprivates/reductions. Patch adds this support by using shared memory
if the total size of the reductions/lastprivates is <= 128 bytes,
then pre-allocated buffer in global memory if size is <= 4K bytes,or
uses malloc/free, otherwise.
Reviewers: gtbercea, kkwli0, grokos
Subscribers: guansong, openmp-commits
Differential Revision: https://reviews.llvm.org/D51875
llvm-svn: 342737
Summary:
Missed operation of the incrementing iterator when required just to
continue execution.
Reviewers: kkwli0, gtbercea, grokos
Subscribers: guansong, openmp-commits
Differential Revision: https://reviews.llvm.org/D51937
llvm-svn: 341964
Some types and callback signatures have changed from TR6 to TR7.
Major changes (only adding signatures and stubs):
(-remove idle callback) done by D48362
-add reduction and dispatch callback
-add get_task_memory and finalize_tool runtime entry points
-ompt_invoker_t becomes ompt_parallel_flag_t
-more types of sync_regions
Patch provided by Simon Convent
Reviewers: hbae, protze.joachim
Differential Revision: https://reviews.llvm.org/D50774
llvm-svn: 341834
Add atomic hint flags to the enum.
The hint parameter type was changed to uint32_t in __kmpc_critical_with_hint()
Patch by Olga Malysheva
Differential Revision: https://reviews.llvm.org/D51235
llvm-svn: 341694
ident flags reserved for atomic hints.
This patch adds omp_sync_hint_t to omp.h and omp_sync_hint_kind to omp_lib.h.
For better maintainability the list of macros for ident flags was replaced with
a enum. The new KMP_IDENT_ATOMIC_HINT_MASK was added to the enum to
support possible future atomic hints.
Also fix omp_lib.h.var to be under 72 chars again after 5.0 OpenMP Memory commit
Patch by Olga Malysheva
Differential Revision: https://reviews.llvm.org/D51233
llvm-svn: 341693
Implemented omp_alloc, omp_free, omp_{set,get}_default_allocator entries,
and OMP_ALLOCATOR environment variable.
Added support for HBW memory on Linux if libmemkind.so library is accessible
(dynamic library only, no support for static libraries).
Only used stable API (hbwmalloc) of the memkind library
though we may consider using experimental API in future.
The ICV def-allocator-var is implemented per implicit task similar to
place-partition-var. In the absence of a requested allocator, the uses the
default allocator.
Predefined allocators (the only ones currently available) are made similar
for C and Fortran, - pointers (long integers) with values 1 to 8.
Patch by Andrey Churbanov
Differential Revision: https://reviews.llvm.org/D51232
llvm-svn: 341687
This is a follow-up to r341371: The new test for PR38704 doesn't
work with Clang 6.0. It uses an UNSUPPORTED: clang-6, but that
hasn't worked because the compiler features weren't known to lit.
llvm-svn: 341448
cuDeviceGetProperties has apparently been deprecated since CUDA 5.0.
Nvidia started using annotations only in CUDA 9.2, so nobody noticed
nor cared before.
The new function returns the same values, tested with a P100.
Differential Revision: https://reviews.llvm.org/D51624
llvm-svn: 341372
* cg and HasCancel in WorkDescr were never read and can be removed.
* This eliminates the last use of priv in ThreadPrivateContext.
* CounterGroup is unused afterwards.
* Remove duplicate external declares in omptarget-nvptx.cu that are
already in the header omptarget-nvptx.h.
Differential Revision: https://reviews.llvm.org/D51622
llvm-svn: 341370
If the runtime is uninitialized the master thread must Enqueue the
state object, and ALL threads must return immediately.
Found post-commit of https://reviews.llvm.org/D51222.
llvm-svn: 341328
Summary:
Implemented simple and lightweight runtime support for SPMD mode-based
constructs. It adds support for L2 sequential parallelism wihtout full
runtime support. Also, patch fixes some use cases for
uninitialized|lightweight runtime.
Reviewers: grokos, kkwli0, Hahnfeld, gtbercea
Subscribers: guansong, openmp-commits
Differential Revision: https://reviews.llvm.org/D51222
llvm-svn: 340944
Summary:
Removed the function that used a lock and varargs
Used the same mechanism as for debug messages
Reviewers: ABataev, gtbercea, grokos, Hahnfeld
Reviewed By: gtbercea, Hahnfeld
Subscribers: mikerice, ABataev, RaviNarayanaswamy, guansong, openmp-commits
Differential Revision: https://reviews.llvm.org/D51226
llvm-svn: 340767
The __kmp_execute_tasks_template() function reads the task_team and
current_task from the thread structure. There appears to be a pathological
timing where the number of threads in the hot team decreases and so a
thread is put in the pool via __kmp_free_thread(). It could be the case that:
1) A thread reads th_task_team into task_team local variables
and is then interrupted by the OS
2) Master frees the thread and sets current task and task team to NULL
3) The thread reads current_task as NULL
When this happens, current_task is dereferenced and a segfault occurs.
This patch just checks for current_task to not be NULL as well.
Differential Revision: https://reviews.llvm.org/D50651
llvm-svn: 340632
If hot teams are not being used, this code could seg fault without the added
check, and does so when composability is used in conjunction with nesting.
The fix prevents the segfault.
Differential Revision: https://reviews.llvm.org/D50649
llvm-svn: 340629
Exclude nested explicit tasks from timing, only outer level explicit task
counted and its time added to barrier arrive time for the thread.
Differential Revision: https://reviews.llvm.org/D50584
llvm-svn: 340628
Summary:
Right now, only the OMP_TARGET_OFFLOAD=DISABLED was implemented. Added support for the other MANDATORY and DEFAULT values.
Reviewers: gtbercea, ABataev, grokos, caomhin, Hahnfeld
Reviewed By: Hahnfeld
Subscribers: protze.joachim, gtbercea, AlexEichenberger, RaviNarayanaswamy, Hahnfeld, guansong, openmp-commits
Differential Revision: https://reviews.llvm.org/D50522
llvm-svn: 340542
The idle callback was removed from the spec as of TR7.
This removes it from the implementation.
Patch provided by Simon Convent
Reviewers: hbae, protze.joachim
Differential Revision: https://reviews.llvm.org/D48362
llvm-svn: 339771
This change fixes an incorrect behavior of the omp_control_tool function when
called from Fortran applications. A tool callback function for this event is
supposed to get NULL for the third argument according to the specification, but
the current implementation just passes a garbage value. A possible fix is to use
the OPTIONAL attribute for the third argument.
Patch by Hansang Bae
Differential Revision: https://reviews.llvm.org/D50565
llvm-svn: 339585
This patch cleans up unused functions, variables, sign compare issues, and
addresses some -Warning flags which are now enabled including -Wcast-qual.
Not all the warning flags in LibompHandleFlags.cmake are enabled, but some
are with this patch.
Some __kmp_gtid_from_* macros in kmp.h are switched to static inline functions
which allows us to remove the awkward definition of KMP_DEBUG_ASSERT() and
KMP_ASSERT() macros which used the comma operator. This had to be done for the
innumerable -Wunused-value warnings related to KMP_DEBUG_ASSERT()
Differential Revision: https://reviews.llvm.org/D49105
llvm-svn: 339393
This patch adds a test using the doacross clauses in OpenMP and removes gcc from
testing kmp_doacross_check.c which is only testing the kmp rather than the
gomp interface.
Differential Revision: https://reviews.llvm.org/D50014
llvm-svn: 338757
This is broken per PR36561 and PR36574, so disable it for now until
somebody interested can take a look. OMPT can still be activated manually
by passing -DLIBOMP_OMPT_SUPPORT=ON during configuration.
Differential Revision: https://reviews.llvm.org/D50086
llvm-svn: 338721
Only supported since GCC 6 and Intel 17.0. However GCC 6.3.0 is
crashing on two of the tests, so disable them as well...
Differential Revision: https://reviews.llvm.org/D50085
llvm-svn: 338720
The taskloop testcase had scheduling effects. Tasks of the taskloop would
sometimes be scheduled before all task were created. The testing is now
split into two phases. First, the task creation on the master is tested,
than the scheduling events of the tasks are tested. Thus, the order of
creation and scheduling events is irrelavant.
Patch by Simon Convent
Reviewed by: protze.joachim, Hahnfeld
Subscribers: openmp-commits
Differential Revision: https://reviews.llvm.org/D50140
llvm-svn: 338580
GCC 4.8.5 defaults to this old C standard. I think we should make the
tests pass a newer -std=c99|c11 but that's too intrusive for now...
Differential Revision: https://reviews.llvm.org/D50084
llvm-svn: 338490
From the bug report, the runtime needs to initialize the nproc variables
(inside middle init) for each root when the task is encountered, otherwise,
a segfault can occur.
Bugzilla: https://bugs.llvm.org/show_bug.cgi?id=36720
Differential Revision: https://reviews.llvm.org/D49996
llvm-svn: 338313
Summary:
When OMPT is not supported the __kmp_omp_task() function is passed the parameters in the wrong order. This is a fix related to patch D47709.
Reviewers: Hahnfeld, sconvent, caomhin, jlpeyton
Reviewed By: Hahnfeld
Subscribers: guansong, openmp-commits
Differential Revision: https://reviews.llvm.org/D50001
llvm-svn: 338295
GCC 8 produces false-positives with this:
In file included from <openmp>/src/runtime/src/kmp_os.h:950,
from <openmp>/src/runtime/src/kmp.h:78,
from <openmp>/src/runtime/src/kmp_environment.cpp:54:
<openmp>/src/runtime/src/kmp_environment.cpp: In function ‘char* __kmp_env_get(const char*)’:
<openmp>/src/runtime/src/kmp_safe_c_api.h:52:50: warning: ‘char* strncpy(char*, const char*, size_t)’ specified bound depends on the length of the source argument [-Wstringop-overflow=]
#define KMP_STRNCPY_S(dst, bsz, src, cnt) strncpy(dst, src, cnt)
~~~~~~~^~~~~~~~~~~~~~~
<openmp>/src/runtime/src/kmp_environment.cpp:97:5: note: in expansion of macro ‘KMP_STRNCPY_S’
KMP_STRNCPY_S(result, len, value, len);
^~~~~~~~~~~~~
<openmp>/src/runtime/src/kmp_environment.cpp:92:28: note: length computed here
size_t len = KMP_STRLEN(value) + 1;
This is stupid because result is allocated with KMP_INTERNAL_MALLOC(len),
so the arguments are correct.
Differential Revision: https://reviews.llvm.org/D49904
llvm-svn: 338283
This change introduces GOMP doacross compatibility. There are 12 new interface
functions 6 for long type and 6 for unsigned long long type:
GOMP_doacross_post, GOMP_doacross_wait, GOMP_loop_doacross_[schedule]_start
where schedule can be static, dynamic, guided, or runtime.
These functions just translate the parameters if necessary and send them
to the corresponding kmp function.
E.g., GOMP_doacross_post() -> __kmpc_doacross_post()
For the GOMP_doacross_post function, there is template specialization to
account for when long is a four byte vs an eight byte type. If it is a
four byte type, then a temporary array has to be created to convert the
four byte integers into eight byte integers and then sending that into
__kmpc_doacross_post(). Because GOMP_doacross_wait uses varargs, it
always needs a temporary array and does not need template specialization.
Differential Revision: https://reviews.llvm.org/D49857
llvm-svn: 338280
This change fixes build errors when building a runtime with adaptive lock stats
enabled. Most of the errors were due to the recent changes in the runtime, but
it seems that we have not tried to build this debug runtime on Windows for a
long time.
Patch by Hansang Bae
Differential Revision: https://reviews.llvm.org/D49823
llvm-svn: 338277
1) Remove unnecessary data from list node structure
2) Remove timerPair in favor of pushing/popping explicitTimers.
This way, nested timers will work properly.
3) Fix #pragma omp critical timers
4) Add histogram capability
5) Add KMP_STATS_FILE formatting capability
6) Have time partitioned into serial & parallel by introducing
partitionedTimers::exchange(). This also counts the number of serial regions
in the executable.
7) Fix up the timers around OMP loops so that scheduling overhead and work are
both counted correctly.
8) Fix up the iterations statistics so they count the number of iterations the
thread receives at each loop scheduling event
9) Change timers so there is only one RDTSC read per event change
10) Fix up the outdated comments for the timers
Differential Revision: https://reviews.llvm.org/D49699
llvm-svn: 338276
Fix the order of callbacks related to the taskloop construct.
Add the iteration_count to work callbacks (according to the spec).
Use kmpc_omp_task() instead of kmp_omp_task() to include OMPT callbacks.
Add a testcase.
Patch by Simon Convent
Reviewed by: protze.joachim, hbae
Subscribers: openmp-commits
Differential Revision: https://reviews.llvm.org/D47709
llvm-svn: 338146
The ompt/tasks/task_types.c testcase did not test untied tasks properly. Now,
frame addresses are tested and two scheduling points are added at which the
task can switch to another thread. Due to scheduling effects, the frame address
could be NULL.
This needed a restructure of the way OMPT callbacks are called.
__ompt_task_finish() now as an extra parameter, whether a task is completed.
Its invocation has been moved into __kmp_task_finish(). Thus, the order of the
writes to the frame addresses is not subject to scheduling effects anymore.
Patch by Simon Convent
Reviewed by: protze.joachim, hbae
Subscribers: openmp-commits
Differential Revision: https://reviews.llvm.org/D49181
llvm-svn: 338145
The two more outputs are needed to match the return addresses when using the
Intel Compiler, as it generates more instructions between the fuzzy-printing
of the address and the runtime call.
Patch by Simon Convent
Reviewed By: protze.joachim, hbae
Differential Revision: https://reviews.llvm.org/D49373
llvm-svn: 338144
This function was not enabled by default and not exported when manually
tweaking the build flags. Additionally it was hard to use since there
is no corresponding __kmp_ft_page_free().
The code itself is questionable because the returned memory address
is padded by an extra pointer which stores the unpadded start of the
allocated region (this would need to be freed).
Differential Revision: https://reviews.llvm.org/D49802
llvm-svn: 338052
The initial commit said that the test passes with Intel Compiler,
so change XFAIL to only list clang and gcc.
Differential Revision: https://reviews.llvm.org/D49801
llvm-svn: 338051
Summary:
1. Fixed internal problem in `__kmpc_barrier` function: SPMD mode
synchronization function should be called only in L1 parallel level.
2. Removed some extra code for synchronization inside of the code, used
`__kmpc_barrier` instead.
3. Some code cleanup.
Reviewers: gtbercea, grokos
Subscribers: openmp-commits
Differential Revision: https://reviews.llvm.org/D49564
llvm-svn: 337691
This change fixes possibly invalid access to the internal data structure during
library shutdown. In a heavily oversubscribed situation, the library shutdown
sequence can reach the point where resources are deallocated while there still
exist threads in their final spinning loop. The added loop in
__kmp_internal_end() checks if there are such busy-waiting threads and blocks
the shutdown sequence if that is the case. Two versions of kmp_wait_template()
are now used to minimize performance impact.
Patch by Hansang Bae
Differential Revision: https://reviews.llvm.org/D49452
llvm-svn: 337486
This patch removes the translation code since this functionality is now implemented in the compiler.
target_data_begin and target_data_end are also patched to handle some special cases that used to be
handled by the obsolete translation function, namely ensure proper alignment of struct members when
we have partially mapped structs. Mapping a struct from a higher address (i.e. not from its beginning)
can result in distortion of the alignment for some of its member fields. Padding restores the original
(proper) alignment.
Differential revision: https://reviews.llvm.org/D44186
llvm-svn: 337455
In revision r336569 (D49036) libomptarget support for multiple nvidia images
has been fixed in case a target region resides inside one or multiple
libraries and in the compiled application. But the issues is still present
for elf images.
This fix will also support multiple images for elf.
Patch by Jannis Klinkenberg
Reviewers: protze.joachim, ABataev, grokos
Reviewed By: protze.joachim, ABataev, grokos
Subscribers: openmp-commits
Differential Revision: https://reviews.llvm.org/D49418
llvm-svn: 337355
Summary:
Should be variable name instead of variable reference. If the variable is
somehow unset, it messes up the if condition expression and causes a CMake
error.
Reviewers: jlpeyton, AndreyChurbanov, Hahnfeld
Reviewed By: Hahnfeld
Subscribers: mgorny, llvm-commits, openmp-commits
Differential Revision: https://reviews.llvm.org/D47221
llvm-svn: 337133
Summary: This patch fixes the data sharing infrastructure to work for the SPMD and non-SPMD cases.
Reviewers: ABataev, grokos, carlo.bertolli, caomhin
Reviewed By: ABataev, grokos
Subscribers: guansong, openmp-commits
Differential Revision: https://reviews.llvm.org/D49204
llvm-svn: 337013
Summary:
Patch fixes the next problems.
1. Removes unused functions from omptarget_nvptx_ThreadPrivateContext
class + simplified data members.
2. Fixed calculation of loop boundaries for dynamic loops with static
scheduling.
3. Introduced saving/restoring of the dynamic loop boundaries to support
several nested parallel dynamic loops.
Reviewers: grokos
Subscribers: guansong, kkwli0, openmp-commits
Differential Revision: https://reviews.llvm.org/D49241
llvm-svn: 336915
This patch introduces the logic implementing hierarchical scheduling.
First and foremost, hierarchical scheduling is off by default
To enable, use -DLIBOMP_USE_HIER_SCHED=On during CMake's configure stage.
This work is based off if the IWOMP paper:
"Workstealing and Nested Parallelism in SMP Systems"
Hierarchical scheduling is the layering of OpenMP schedules for different layers
of the memory hierarchy. One can have multiple layers between the threads and
the global iterations space. The threads will go up the hierarchy to grab
iterations, using possibly a different schedule & chunk for each layer.
[ Global iteration space (0-999) ]
(use static)
[ L1 | L1 | L1 | L1 ]
(use dynamic,1)
[ T0 T1 | T2 T3 | T4 T5 | T6 T7 ]
In the example shown above, there are 8 threads and 4 L1 caches begin targeted.
If the topology indicates that there are two threads per core, then two
consecutive threads will share the data of one L1 cache unit. This example
would have the iteration space (0-999) split statically across the four L1
caches (so the first L1 would get (0-249), the second would get (250-499), etc).
Then the threads will use a dynamic,1 schedule to grab iterations from the L1
cache units. There are currently four supported layers: L1, L2, L3, NUMA
OMP_SCHEDULE can now read a hierarchical schedule with this syntax:
OMP_SCHEDULE='EXPERIMENTAL LAYER,SCHED[,CHUNK][:LAYER,SCHED[,CHUNK]...]:SCHED,CHUNK
And OMP_SCHEDULE can still read the normal SCHED,CHUNK syntax from before
I've kept most of the hierarchical scheduling logic inside kmp_dispatch_hier.h
to try to keep it separate from the rest of the code.
Differential Revision: https://reviews.llvm.org/D47962
llvm-svn: 336571
Summary:
Currently Cuda plugin supports loading of the single image, though we
may have the executable with the several images, if it has target
regions inside of the dynamically loaded library. Patch allows to load
multiple images.
Reviewers: grokos
Subscribers: guansong, openmp-commits, kkwli0
Differential Revision: https://reviews.llvm.org/D49036
llvm-svn: 336569
This patch reorganizes the loop scheduling code in order to allow hierarchical
scheduling to use it more effectively. In particular, the goal of this patch
is to separate the algorithmic parts of the scheduling from the thread
logistics code.
Moves declarations & structures to kmp_dispatch.h for easier access in
other files. Extracts the algorithmic part of __kmp_dispatch_init() and
__kmp_dispatch_next() into __kmp_dispatch_init_algorithm() and
__kmp_dispatch_next_algorithm(). The thread bookkeeping logic is still kept in
__kmp_dispatch_init() and __kmp_dispatch_next(). This is done because the
hierarchical scheduler needs to access the scheduling logic without the
bookkeeping logic. To prepare for new pointer in dispatch_private_info_t, a
new flags variable is created which stores the ordered and nomerge flags instead
of them being in two separate variables. This will keep the
dispatch_private_info_t structure the same size.
Differential Revision: https://reviews.llvm.org/D47961
llvm-svn: 336568
These are preliminary changes that attempt to use C++11 Atomics in the runtime.
We are expecting better portability with this change across architectures/OSes.
Here is the summary of the changes.
Most variables that need synchronization operation were converted to generic
atomic variables (std::atomic<T>). Variables that are updated with combined CAS
are packed into a single atomic variable, and partial read/write is done
through unpacking/packing
Patch by Hansang Bae
Differential Revision: https://reviews.llvm.org/D47903
llvm-svn: 336563
The flag "--no-as-needed" is not recognized by the linker on macOS making the following tests fail:
ompt/loadtool/tool_available/tool_available.c
ompt/loadtool/tool_not_available/tool_not_available.c
This patch removes this flag for macOS and adds it only for Linux and Windows.
I tested it on Ubuntu 16.04 and macOS HighSierra, with Clang/LLVM 6.0.1 and OpenMP trunk.
This solution was also discussed in the OpenMP-dev mailing list.
Patch provided by Simone Atzeni
Differential Revision: https://reviews.llvm.org/D48888
llvm-svn: 336327
The testcase potentially fails when a thread is reused.
The added synchronization makes sure this does not happen.
Patch provided by Simon Convent
Differential Revision: https://reviews.llvm.org/D48932
llvm-svn: 336326
When compiling with icc, there is a problem with reenter frame addresses in
parallel_begin callbacks in the interoperability.c testcase. (The address is
not available. thus NULL)
Using alloca() forces availability of the frame pointer.
Patch provided by Simon Convent
Differential Revision: https://reviews.llvm.org/D48282
llvm-svn: 336088
Several runtime entry points have not been tested from non-OpenMP threads. This
adds tests to an existing testcase. While at it, the testcase was reformatted
Patch provided by Simon Convent
Differential Revision: https://reviews.llvm.org/D48124
llvm-svn: 336087
Especially the thread_end callback has not been tested before.
This adds a testcase for nested and non-nested threads.
Patch provided by Simon Convent
Differential Revision: https://reviews.llvm.org/D47824
llvm-svn: 336086
The current implementation always provides the thread-num for the current
parallel region. This patch fixes the behavior for ancestor levels >0.
Differential Revision: https://reviews.llvm.org/D46533
llvm-svn: 336085
Summary:
Patch fixes several problems in the implementation of NVPTX RTL.
1. Detection of the last iteration for loops with static scheduling, no chunks.
2. Fixes reductions for the serialized parallel constructs.
3. Fixes handling of the barriers.
Reviewers: grokos
Reviewed By: grokos
Subscribers: Hahnfeld, guansong, openmp-commits
Differential Revision: https://reviews.llvm.org/D48480
llvm-svn: 335469
Upcoming changes to FileCheck will modify CHECK-DAG to not match
overlapping regions of the input. This test was found to be affected
because it expects to find four threads to invoke events of type
ompt_event_implicit_task_begin. It turns out this is wrong because
OMP_THREAD_LIMIT is set to 2, so there are only two threads. The
rest of the test got it right so it went unnoticed until now.
(Rewrite test and apply clang-format to it as discussed in the past.)
Differential Revision: https://reviews.llvm.org/D47119
llvm-svn: 333361
The generic entry points for static loop scheduling previously
hardcoded that the runtime was initialized. This can be wrong if
the compiler analyzes that the runtime is not needed and calls
the init functions accordingly.
This didn't affect clang-ykt because they have entry points for
different combinations of SPMD x Runtime not needed. I didn't do
measurements yet but with inlining we might get away with always
calling the generic interface and letting compiler and runtime
figure out the rest.
In any case, a correct runtime is always better than having
functions that may only be called if previous calls passed in
a specific set of arguments!
Differential Revision: https://reviews.llvm.org/D47131
llvm-svn: 333285
Introduce OPENMP_INSTALL_LIBDIR and use in all install() commands.
This also fixes installation of libomptarget-nvptx that previously
didn't honor {OPENMP,LLVM}_LIBDIR_SUFFIX.
Differential Revision: https://reviews.llvm.org/D47130
llvm-svn: 333284
The existing implementation of the dynamic scheduling
breaks the contract introduced by the original openmp
runtime and, thus, is incorrect. Patch fixes it and
introduces correct dynamic scheduling model.
Thanks to Alexey Bataev for submitting this patch.
Differential Revision: https://reviews.llvm.org/D47333
llvm-svn: 333225
We already know where the CUDA SDK is, so there is no point in
letting Clang search for it again and possibly finding no or
a different installation.
--cuda-path is supported since the beginning of CUDA support in
Clang, so making this required doesn't impose additional restrictions.
Differential Revision: https://reviews.llvm.org/D46930
llvm-svn: 332495
Move all logic related to selecting the bitcode compiler and linker
into a new file and dynamically test required compiler flags. This
also adds -fcuda-rdc for Clang trunk as previously attempted in D44992
which fixes the build.
As a result this change also enables building the library by default
if all prerequisites are met.
Differential Revision: https://reviews.llvm.org/D46901
llvm-svn: 332494
Summary: Add function to the NVPTX libomptarget library that will return true if the current target region is being executed in SPMD mode.
Reviewers: ABataev, grokos, carlo.bertolli, caomhin
Reviewed By: grokos
Subscribers: guansong, openmp-commits
Differential Revision: https://reviews.llvm.org/D46840
llvm-svn: 332360
implicit_task_end callbacks in nested parallel regions did not always give the
correct thread_num, since the inner parallel region may have already been
finalized.
Now, the thread_num is stored at the beginning of the implicit task and
retrieved at the end, whenever necessary.
A testcase was added as well.
Differential Revision: https://reviews.llvm.org/D46260
llvm-svn: 331632
The api_calls_misc.c testcase tests the following api calls:
ompt_get_callback()
ompt_get_state()
ompt_enumerate_states()
ompt_enumerate_mutex_impls()
These have not been tested previously.
The api_calls.c testcase has been renamed to api_calls_places.c because it only tests api calls that are related to places.
Differential Revision: https://reviews.llvm.org/D42523
llvm-svn: 331631
Summary:
Enable the device side debug messages at compile time, use env var to control at runtime.
To achieve this, an environment data block is passed to the device lib when it is loaded.
By default, the message is off, to enable it, a user need to set LIBOMPDEVICE_DEBUG=1.
Reviewers: grokos
Reviewed By: grokos
Subscribers: openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D46210
llvm-svn: 331550
Summary: Minor printf format correction. NVCC ignore those. Clang will give warning on these if debug is enabled.
Reviewers: grokos
Reviewed By: grokos
Subscribers: openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D45528
llvm-svn: 330944
Summary: The LIBOMPTARGET_NVPTX_DEBUG flag is inconsistent between using nvcc to generate .a file and clang to generate .bc file. Sync the two setting so we can get debug messages from the bc file path as well.
Reviewers: grokos
Subscribers: Hahnfeld, openmp-commits, mgorny
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D45530
llvm-svn: 330477
Currently, the affinity API reports garbage for the initial place list and any
thread's place lists when using KMP_AFFINITY=none|compact|scatter.
This patch does two things:
for KMP_AFFINITY=none, Creates a one entry table for the places, this way, the
initial place list is just a single place with all the proc ids in it. We also
set the initial place of any thread to 0 instead of KMP_PLACE_ALL so that the
thread reports that single place (place 0) instead of garbage (-1) when using
the affinity API.
When non-OMP_PROC_BIND affinity is used
(including KMP_AFFINITY=compact|scatter), a thread's place list is populated
correctly. We assume that each thread is assigned to a single place. This is
implemented in two of the affinity API functions
Differential Revision: https://reviews.llvm.org/D45527
llvm-svn: 330283
This patch introduces GOMP_taskloop to our API. It adds GOMP_4.5 to our
version symbols. Being a wrapper around __kmpc_taskloop, the function
creates a task with the loop bounds properly nested in the shareds so that
the GOMP task thunk will work properly. Also, the firstprivate copy constructors
are properly handled using the __kmp_gomp_task_dup() auxiliary function.
Currently, only linear spawning of tasks is supported
for the GOMP_taskloop interface.
Differential Revision: https://reviews.llvm.org/D45327
llvm-svn: 330282
Summary:
This one line change is to remove this warning message
"warning: integer conversion resulted in a change of sign"
Reviewers: grokos
Reviewed By: grokos
Subscribers: openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D45415
llvm-svn: 329713
This change removes the unnecessary lock operation on __kmp_initz_lock inside
the __kmp_atfork_child() function for Linux; the lock variable is initialized
in the same function later.
Patch by Hansang Bae
Differential Revision: https://reviews.llvm.org/D44949
llvm-svn: 328900
The summarizeStats.py script processes raw data provided by the
instrumented (stats-gathering) OpenMP* runtime library. It provides:
1) A radar chart which plots counters as frequency (per GigaTick) of use within
the program. The frequencies are plotted as log10, however values less than
one are kept as it is and represented in red color. This was done to help
visualize the differences better.
2) Pie charts separating total time as compute and non-compute. The compute and
non-compute times have their own pie charts showing the constructs that
contributed to them. The percentages listed are with respect to the total
time.
3) '.csv' file with percentage of time spent within the different constructs.
The script can be used as:
$ python $PATH_TO_SCRIPT/summarizeStats.py instrumented1.csv instrumented2.csv
Patch by Taru Doodi
Differential Revision: https://reviews.llvm.org/D41838
llvm-svn: 328568
Summary: The global stack initialization function may be called multiple times. The initialization of the shared memory slots should only happen when the function is called for the first time for a given warp master thread.
Reviewers: grokos, carlo.bertolli, ABataev, caomhin
Reviewed By: grokos
Subscribers: guansong, openmp-commits
Differential Revision: https://reviews.llvm.org/D44754
llvm-svn: 328148
Summary: The check for the master warp must take into consideration the actual number of warps: the master warp is equal to the last active warp not necessarily WARPSIZE - 1.
Reviewers: grokos, carlo.bertolli, ABataev, caomhin
Reviewed By: grokos
Subscribers: guansong, openmp-commits
Differential Revision: https://reviews.llvm.org/D44537
llvm-svn: 328146
Summary:
This patch allows worker to have a global memory stack managed by the runtime. This patch is needed for completeness and consistency with the globalization policy: if a worker-side variable escapes the current context it then needs to be globalized.
Until now, only the master thread was allowed to have such a stack. These global values can now potentially be shared amongst workers if the semantics of the OpenMP program require it.
Reviewers: ABataev, grokos, carlo.bertolli, caomhin
Reviewed By: grokos
Subscribers: guansong, openmp-commits
Differential Revision: https://reviews.llvm.org/D44487
llvm-svn: 328144
Added settings code to read OMP_TARGET_OFFLOAD environment variable. Added
target-offload-var ICV as __kmp_target_offload, set via OMP_TARGET_OFFLOAD,
if available, otherwise defaulting to DEFAULT. Valid values for the ICV are
specified as enum values {0,1,2} for disabled, default, and mandatory. An
internal API access function __kmpc_get_target_offload is provided.
Patch by Terry Wilmarth
Differential Revision: https://reviews.llvm.org/D44577
llvm-svn: 328046
Summary:
Allow the runtime to use the existing shared memory statically allocated slots.
When a variable is globalized, the underlying memory can be either shared or global memory (both have block-wide visibility). In this case, we allow that the storage to use a limited amount of shared memory that has been statically allocated already. Only if shared memory doesn't prove to be enough do we then invoke malloc() to create a new global memory slot.
Reviewers: ABataev, carlo.bertolli, grokos, caomhin
Reviewed By: grokos
Subscribers: guansong, openmp-commits
Differential Revision: https://reviews.llvm.org/D44486
llvm-svn: 327639
Summary: To save on calls to malloc, this patch enables the re-use of pre-allocated global memory slots.
Reviewers: ABataev, grokos, carlo.bertolli, caomhin
Reviewed By: grokos
Subscribers: guansong, openmp-commits
Differential Revision: https://reviews.llvm.org/D44470
llvm-svn: 327637
Summary:
This patch adds support for the sharing of variables from the master thread of a team to the worker threads of the team.
The runtime uses a stack structure implemented as a doubly-linked list of slots with each slot having the exact same size as the size requested. This implementation leverages existing data structures. The runtime functions are added as separate functions to avoid interfering with the current interface.
Limitations to be addressed in future patches:
- This current patch only employs global memory. In a future patch we will enable to usage for shared memory as an optimization.
- Allow the allocation of several requested sizes in the same slot.
Reviewers: ABataev, grokos, caomhin, carlo.bertolli
Reviewed By: grokos
Subscribers: Hahnfeld, guansong, openmp-commits
Differential Revision: https://reviews.llvm.org/D44260
llvm-svn: 327440
Summary: To make the two parts of the union have the same size, the size of vect needs to be increased by 16 bits.
Reviewers: grokos, carlo.bertolli, caomhin, ABataev
Reviewed By: grokos, ABataev
Subscribers: fedor.sergeev, guansong, openmp-commits
Differential Revision: https://reviews.llvm.org/D44254
llvm-svn: 327040
Summary:
This patch reverts the changes to libomptarget that were coupled with the changes to Clang code gen for data sharing using shared memory. A similar patch exists for Clang: D43625
Shared memory is meant to be used as an optimization on top of a more general scheme. So far we didn't have a global memory implementation ready so shared memory was a solution which applied to the current level of OpenMP complexity supported by trunk on GPU devices (due to the missing NVPTX backend patch this functionality has never been exercised). Now that we have a global memory solution this patch is "in the way" and needs to be removed (for now). This patch (or an equivalent version of it) will be put out for review once the global memory scheme is in place.
Reviewers: ABataev, grokos, carlo.bertolli, caomhin
Reviewed By: grokos
Subscribers: Hahnfeld, guansong, openmp-commits
Differential Revision: https://reviews.llvm.org/D43626
llvm-svn: 326950
We have to ensure that the runtime is initialized _before_ waiting
for the two started threads to guarantee that the master threads
post their ompt_event_thread_begin before the worker threads. This
is not guaranteed in the parallel region where one worker thread
could start before the other master thread has invoked the callback.
The problem did not happen with Clang becauses the generated code
calls __kmpc_global_thread_num() and cashes its result for functions
that contain OpenMP pragmas.
Differential Revision: https://reviews.llvm.org/D43882
llvm-svn: 326435
The thread_num parameter of ompt_get_task_info() was not being used previously,
but need to be set.
The print_task_type() function (form the task-types.c testcase) was merged into
the print_ids() function (in callback.h). Testing of ompt_get_task_info() was
added to the task-types.c testcase. It was not tested extensively previously.
Differential Revision: https://reviews.llvm.org/D42472
llvm-svn: 326338
The main change of this patch is to insert {{.*}} in current_address=[[RETURN_ADDRESS_END]].
This is needed to match any of the alternatively printed addresses.
Additionally, clang-format is applied to the two tests.
Differential Revision: https://reviews.llvm.org/D43115
llvm-svn: 326312
This is required to be NULL for implicit barriers at the end of a
parallel region. Noticed in review of D43191.
Differential Revision: https://reviews.llvm.org/D43308
llvm-svn: 325922
The compiler inlines the user code in the task. Check for that case at
runtime by comparing the frame addresses and print the expected exit
address.
Also showcase how I think the OMPT tests could be reformatted to match
LLVM's code style. In my opinion it would be great to that kind of change
to all tests that need to be touched for whatever reason...
Differential Revision: https://reviews.llvm.org/D43191
llvm-svn: 325921
Test whether OMPT-callbacks for two threads that initiate a parallel region are correct.
Differential Revision: https://reviews.llvm.org/D41942
llvm-svn: 325423
Only use ompt_ functions when testing OMPT in api_calls testcase.
Add size parameter to print_list.
Fix small bug in implementation of ompt_get_partition_place_nums(): return correct length.
Differential Revision: https://reviews.llvm.org/D42162
llvm-svn: 325422
GlobalISel doesn't yet implement blockaddress and falls back to
SelectionDAG. This results in additional branch instruction to
the next basic block which breaks the OMPT tests.
Disable GlobalISel for now when compiling the tests because fixing
them is not easily possible. See http://llvm.org/PR36313 for full
discussion history.
Differential Revision: https://reviews.llvm.org/D43195
llvm-svn: 325218
This affects all outlined functions, not just tasks! Only show warning
when using Clang 5.0 or later.
Differential Revision: https://reviews.llvm.org/D43190
llvm-svn: 325131
Summary:
Different NVIDIA GPUs support different compute capabilities. To enable the inlining of runtime functions and the best performance on different generations of NVIDIA GPUs, a bc library for each compute capability needs to be compiled. The same compiler build will then be usable in conjunction with multiple generations of NVIDIA GPUs.
To differentiate between versions of the same bc lib, the output file name will contain the compute capability ID.
Depends on D14254
Reviewers: Hahnfeld, hfinkel, carlo.bertolli, caomhin, ABataev, grokos
Reviewed By: Hahnfeld, grokos
Subscribers: guansong, mgorny, openmp-commits
Differential Revision: https://reviews.llvm.org/D41724
llvm-svn: 324904
Tests the search for tools as defined in the spec. The OMP_TOOL_LIBRARIES
environment variable contains paths to the following files(in that order)
-to a nonexisting file
-to a shared library that does not have a ompt_start_tool function
-to a shared library that has an ompt_start_tool implementation returning NULL
-to a shared library that has an ompt_start_tool implementation returning a
pointer to a valid instance of ompt_start_tool_result_t
The expected result is that the last tool gets active and can print in the
thread-begin callback.
Differential Revision: https://reviews.llvm.org/D42166
llvm-svn: 324588
Add a testcase that checks wheter the runtime can handle an ompt_start_tool
method that returns NULL indicating that no tool shall be loaded.
All tool_available testcases need a separate folder to avoid file conflicts for
the generated tools.
Differential Revision: https://reviews.llvm.org/D41904
llvm-svn: 324587
Summary: This patch extends the libomptarget functionality in patch D14254 with support for the data sharing scheme for supporting implicitly shared variables. The runtime therefore maintains a list of references to shared variables.
Reviewers: carlo.bertolli, ABataev, Hahnfeld, grokos, caomhin, hfinkel
Reviewed By: Hahnfeld, grokos
Subscribers: guansong, llvm-commits, openmp-commits
Differential Revision: https://reviews.llvm.org/D41485
llvm-svn: 324495
If tool initialization returns 0, OMPT should not be active. The current
implementation provided some callback invocations in this case.
Differential Revision: https://reviews.llvm.org/D42709
llvm-svn: 324320
https://reviews.llvm.org/D42757
The method ThreadsInTeam is used to determine the number of threads to be used in a parallel region under SPMD mode (see line 127 of supporti.h in libomptarget/deviceRTLs/nvptx/src/). This patch fixes the corresponding debug print upon initialization of the kernel in SPMD mode.
llvm-svn: 323978
That's what we really need to link the CUDA plugin against,
not the CUDA runtime API in CUDA_LIBRARIES! While the latter
comes with the CUDA SDK, the Driver API is installed with
the kernel driver and there is at most one per system. As
fallback we can use the stubs library distributed with the
CUDA SDK for linking.
Differential Revision: https://reviews.llvm.org/D42643
llvm-svn: 323787
Use equivalents for the last calls to the Runtime API. Remove
stray assert in case of an error found during review, we should
only return OFFLOAD_FAIL.
Differential Revision: https://reviews.llvm.org/D42686
llvm-svn: 323786
This patch implements the device runtime library whose interface is used in the code generation for OpenMP offloading devices.
Currently there is a single device RTL written in CUDA meant to CUDA enabled GPUs.
The interface is a variation of the kmpc interface that includes some extra calls to do thread and storage management that only make sense for a GPU target.
Differential revision: https://reviews.llvm.org/D14254
llvm-svn: 323649
Use fuzzy return addresses in lock testcases so that these
testcases can also be run using the Intel Compiler.
Patch by Simon Convent!
Differential Revision: https://reviews.llvm.org/D41896
llvm-svn: 323529
Add Workaround for Intel Compiler Bug with Case#: 03138964
A critical region within a nested task causes a segfault in icc 14-18:
int main()
{
#pragma omp parallel num_threads(2)
#pragma omp master
#pragma omp task
#pragma omp task
#pragma omp critical
printf("test\n");
}
When the critical region is in a separate function, the segault does not occur.
So we add noinline to make sure that the function call stays there.
Differential Revision: https://reviews.llvm.org/D41182
llvm-svn: 322622
The defintion is not part of the spec and thus should not have the prefix
"ompt_" but rather a prefix that indicates that this is implementation
specific.
Differential Revision: https://reviews.llvm.org/D41166
llvm-svn: 322621
When the current thread is not an (initialized) OpenMP thread, the runtime
entry points return values that correspond to "not available" or similar
Differential Revision: https://reviews.llvm.org/D41167
llvm-svn: 322620
If user requested affinity with granularity=tile we need to either use HWLOC
or ignore the request. The change allows user to not specify
KMP_TOPOLOGY_METHOD=hwloc and choose it automatically instead.
Patch by Andrey Churbanov
Differential Revision: https://reviews.llvm.org/D40905
llvm-svn: 322205
This change simplifies __kmp_expand_threads to take a single argument.
Previously, it allowed two arguments and had logic to decide on different
potential expansion sizes. However, no calls to __kmp_expand_threads in the
runtime make use of this extra logic. Thus the extra argument and logic is
removed here.
Patch by Terry Wilmarth
Differential Revision: https://reviews.llvm.org/D41836
llvm-svn: 322204
This change improves stability of the runtime when the application forks child
processes. Acquiring/releasing __kmp_initz_lock and __kmp_forkjoin_lock in the
atfork handlers insures that the actual fork does not occur while those two
locks are held, and __kmp_itt_reset() reverts the itt's global state to the
initial state which also initializes the mutex stored in the global state.
Some missing initialization code was also inserted in the child's atfork handler.
Patch by Hansang Bae
Differential Revision: https://reviews.llvm.org/D41462
llvm-svn: 322202
This field is defined as kmp_int32, so we should use neither
pointers to kmp_int64 nor 64 bit atomic instructions.
(Found while testing on a Raspberry Pi, 32 bit ARM)
Differential Revision: https://reviews.llvm.org/D41656
llvm-svn: 321964
This patch enables OMPT by default if version 50 or later is built and the config says, that OMPT will be supported.
Differential Revision: https://reviews.llvm.org/D41508
llvm-svn: 321675
We now have several options that apply for both libraries and they
shouldn't be documented in multiple files. When already merging
the two Build_With_CMake.txt documents, convert them to
reStructuredText which is used for all of LLVM's documentation.
Differential Revision: https://reviews.llvm.org/D40920
llvm-svn: 321481
As for normal task creation, the task frame addresses need to be stored
for the encountering task.
Differential Revision: https://reviews.llvm.org/D41165
llvm-svn: 321421
The compiler warns that _BSD_SOURCE is deprecated and _DEFAULT_SOURCE should
be used instead. We keep _BSD_SOURCE for older compilers, that don't know
about _DEFAULT_SOURCE.
The linker drops the tool when linking, since there is no visible need for
the library. So we need to tell the linker, that the tool should be linked
anyway.
Differential Revision: https://reviews.llvm.org/D41499
llvm-svn: 321362
The format string for hints only prints the second argument (string) and drops
the first argument (hint id). Depending on how you read the POSIX text for
printf, this could be valid. But for practical reason, i.e., unpacking the
va_list passed to printf based on the formating information, it makes sense
to fix the implementation and not pass the id for hint.
Failing testcases were:
misc_bugs/teams-reduction.c
ompt/parallel/not_enough_threads.c
Differential Revision: https://reviews.llvm.org/D41504
llvm-svn: 321361
This function is defined in OpenMP-TR6 section 4.1.5.1.6
The functions was not implemented yet.
Since ompt-functions can only be called after the runtime was initialized and
has loaded a tool, it can assume the runtime to be initialized. In contrast
to omp_get_num_procs which needs to check whether the runtime is initialized.
Differential Revision: https://reviews.llvm.org/D40949
llvm-svn: 321269
This revision fixes failing testcases with parallel for loops and the gomp
interface. The return address needs to be stored at entry to runtime.
The storage is cleared on usage, so we need to update the storage before
calling again internal functions, that will trigger event callbacks.
Differential Revision: https://reviews.llvm.org/D41181
llvm-svn: 321265
We use the bitmap ompt_enabled thoughout the runtime, to avoid loading the
vector of callback functions when testing if specific code should be executed.
Before invoking an event callback function, the pointer is tested for NULL.
This revision resets the corresponding bit in ompt_enabled to 0 if
NULL is passed in set_callback.
Differential Revision: https://reviews.llvm.org/D41171
llvm-svn: 321264
Clang 5 or higher adds an intermediate function call in certain cases when
compiling with debug flag. This revision updates the testcases to work
correctly.
Differential Revision: https://reviews.llvm.org/D40595
llvm-svn: 321263
Reasons for expected failures are mainly bugs when using lables in OpenMP regions
or missing support of some OpenMP features.
For some worksharing clauses, support to distinguish the kind of workshare was
added just recently.
If an issue was fixed in a minor release version of a compiler, we flag the
test as unsupported for this compiler version to avoid false positives.
Same for fixes that where backported to older compiler versions.
Differential Revision: https://reviews.llvm.org/D40384
llvm-svn: 321262
There are two /proc/cpuinfo layots in use for AArch64: old and new.
The old one has all 'processor : n' lines in one section, hence
checking for duplications does not make sense.
Differential Revision: https://reviews.llvm.org/D41000
llvm-svn: 320593
All architectures except x86_64 used the linear barrier implementation
by default which doesn't give good performance for a larger number
of threads.
Improvements for PARALLEL overhead (EPCC) with this patch on a Power8
system (2 sockets x 10 cores x 8 threads, OMP_PLACES=cores)
20 threads: 4.55us -> 3.49us
40 threads: 8.84us -> 4.06us
80 threads: 19.18us -> 4.74us
160 threads: 54.22us -> 6.73us
Differential Revision: https://reviews.llvm.org/D40358
llvm-svn: 320152
To make thread affinity work according to the OpenMP spec, the
runtime needs information about the hardware topology. On Linux
the default way is to parse /proc/cpuinfo which contains this
information for x86 machines but (at least) not for AArch64 and
Power architectures.
Fortunately, there is a different code path which is able to get
that data from sysfs. The needed patch has landed in 2006 for
Linux 2.6.16 which is safe to assume nowadays (even RHEL 5 had
a kernel version derived from 2.6.18, and we are now at RHEL 7!).
Differential Revision: https://reviews.llvm.org/D40357
llvm-svn: 320151
Otherwise I see hangs in the omp_single_copyprivate test when
compiling in release mode. With the debug assertions, I get a
failure `head > 0 && tail > 0`.
Differential Revision: https://reviews.llvm.org/D40722
llvm-svn: 320150
This last of four patches adds a new file for the interface
functions that Clang uses during code generation. The only
change except simply moving the current code is renaming the
function CheckDeviceAndCtors() and using the correct type for
64bit device ids.
Differential Revision: https://reviews.llvm.org/D40801
llvm-svn: 319972
This third patch moves the implementation of the user-facing
OpenMP API functions into its own file. For now, the code is
only moved, no cleanups applied yet.
Differential Revision: https://reviews.llvm.org/D40800
llvm-svn: 319971
This is the second patch to split the current monolithic
implementation into separate files. Note that this change
doesn't cleanup the code yet.
Differential Revision: https://reviews.llvm.org/D40799
llvm-svn: 319970
This is the first of four patches to split the target agnostic
library into multiple (smaller) files. It only moves the code
to separate implementation files and does no cleanup (yet) except
removing unneeded headers.
Differential Revision: https://reviews.llvm.org/D40798
llvm-svn: 319969
Future patches will add (private) header files in src/ that should
not be visible to plugins, so move the "public" ones to a new
include/ directory. This is still internal in a sense that the
contained files won't be installed for the user.
Similarly, the target agnostic offloading library should be built
directly in src/. The parent directory is responsible for finding
dependencies and including all subdirectories.
Differential Revision: https://reviews.llvm.org/D40797
llvm-svn: 319968
Redundant extra verbose output of binding to full mask in case
affinity=balanced or OMP_PLACES=<any> or OMP_PROC_BIND=<any>
Differential Revision: https://reviews.llvm.org/D40624
llvm-svn: 319960
This change is a trivial fix for enums that removes specification of "last" or
"upper" values, or other boundary values. This simplifies the code in places,
and results in never needing to update the "upper" values again.
Patch by Terry Wilmarth
Differential Revision: https://reviews.llvm.org/D40804
llvm-svn: 319957
The runtime will use the global kmp_critical_name as a lock and
tries to atomically store a pointer in there. This will fail
if the global is only aligned by 4 bytes, the size of one int32_t
element. Use a union to ensure the global is aligned to the size
of a pointer on the current platform.
llvm-svn: 319811
__kmpc_reduce_nowait() correctly swapped the teams for reductions
in a teams construct. Apply the same logic to __kmpc_reduce() and
__kmpc_reduce_end().
Differential Revision: https://reviews.llvm.org/D40753
llvm-svn: 319788
Perform a nested CMake invocation to avoid writing our own parser
for compiler versions when we are not testing the in-tree compiler.
Use the extracted information to mark a test as unsupported that
hangs with Clang prior to version 4.0.1 and restrict tests for
libomptarget to Clang version 6.0.0 and later.
Differential Revision: https://reviews.llvm.org/D40083
llvm-svn: 319448
This change makes kmp_r_sched_t type into a union for simpler
comparisons and assignments
Patch by Terry Wilmarth
Differential Revision: https://reviews.llvm.org/D40374
llvm-svn: 319379
kmp_aligned_malloc() always returned NULL on Windows (stub library only)
that may cause Fortran application crash. With this change all memory
allocation functions were fixed to use aligned{m,re,rec}alloc() to
allocate/reallocate memory. To deallocate that memory _aligned_free() is
used in kmp_free().
Patch by Olga Malysheva
Differential Revision: https://reviews.llvm.org/D40296
llvm-svn: 319375
Added two warnings:
1) Before building the topology map check if tiles are requested but the
topo method is not hwloc;
2) After building the topology map check if tiles are requested but not
detected by the library.
Patch by Olga Malysheva
Differential Revision: https://reviews.llvm.org/D40340
llvm-svn: 319374
Fortran array elements made default integer in OMP_GET_PLACE_PROC_IDS and
OMP_GET_PARTITION_PLACE_NUMS subroutines, otherwise call to them produces
incorrect result.
Patch by Olga Malysheva
Differential Revision: https://reviews.llvm.org/D40356
llvm-svn: 319372
The code for the two OpenMP runtime libraries was very similar.
Move to common CMake file that is included and provides a simple
interface for adding testsuites. Also add a common check-openmp
target that runs all testsuites that have been registered.
Note that this renames all test options to the common OPENMP
namespace, for example OPENMP_TEST_C_COMPILER instead of
LIBOMP_TEST_COMPILER and so on.
Differential Revision: https://reviews.llvm.org/D40082
llvm-svn: 319343
These are needed by both libraries, so we can do that in a
common namespace and unify configuration parameters.
Also make sure that the user isn't requesting libomptarget
if the library cannot be built on the system. Issue an error
in that case.
Differential Revision: https://reviews.llvm.org/D40081
llvm-svn: 319342
As a first step, this allows us to generalize the detection of
standalone builds and make it fully compatible when building in
llvm/runtimes/ which automatically sets OPENMP_STANDLONE_BUILD.
Differential Revision: https://reviews.llvm.org/D40080
llvm-svn: 319341
Summary:
We want to automatically copy the appropriate mailing list
for review requests to the openmp repository.
For context, see the proposal and discussion here:
http://lists.llvm.org/pipermail/cfe-dev/2017-November/056032.html
Similar to D40179, I set up a new Diffusion repository with callsign
"OMP" for OpenMP:
https://reviews.llvm.org/source/openmp/
This explicitly updates openmp's .arcconfig to point to the new
OMP repository in Diffusion, which will let us use Herald rule H272
to automatically subscribe openmp-commits to review requests.
Reviewers: hans, grokos, Hahnfeld
Reviewed By: grokos
Subscribers: sammccall, klimek, openmp-commits
Differential Revision: https://reviews.llvm.org/D40499
llvm-svn: 319254
Power has a weak consistency model so we need memory barriers to
make writes (both from runtime and from user code) available for
all threads.
Differential Revision: https://reviews.llvm.org/D40175
llvm-svn: 318848
We have just fixed the codegen of omp_is_initial_device() to reliably work
when offloading to the same device, see commit r316001. This fixes the
failing tests that were the reason why we disabled the library for 5.0.
Differential Revision: https://reviews.llvm.org/D39052
llvm-svn: 318847
This is the libomptarget-side patch which changes the __tgt_* API function signatures in preparation for the new map interface.
Changes are: Device IDs 32bits --> 64bits, Flags 32bits --> 64bits
Differential revision: https://reviews.llvm.org/D40313
llvm-svn: 318790
These tests were failing rarely on my MacBook when there was some
activity in the background. Read: one of a thousand executions?
* sections.c missed the sorting based on thread ids. This worked
as long as the master thread finished its section before the
worker thread started the second one but failed if the master
thread was put to sleep by the OS.
* The checks in single.c assumed that the master thread executes
the single region which works most of the time because it is
usually faster than the newly spawned worker thread.
Differential Revision: https://reviews.llvm.org/D39853
llvm-svn: 318527
The testsuite directory is not used or updated and confuses new users to the
OpenMP project. These tests were rewritten using the lit format and put under
the runtime/test directory. This patch removes the entire testsuite/ directory.
Differential Revision: https://reviews.llvm.org/D39767
llvm-svn: 318056
Traditionally, the library had a weak symbol for ompt_start_tool()
that served as fallback and disabled OMPT if called. Tools could
provide their own version and replace the default implementation
to register callbacks and lookup functions. This mechanism has
worked reasonably well on Linux systems where this interface was
initially developed.
On Darwin / Mac OS X the situation is a bit more complicated and
the weak symbol doesn't work out-of-the-box. In my tests, the
library with the tool needed to link against the OpenMP runtime
to make the process work. This would effectively mean that a tool
needed to choose a runtime library whereas one design goal of the
interface was to allow tools that are agnostic of the runtime.
The solution is to use dlsym() with the argument RTLD_DEFAULT so
that static implementations of ompt_start_tool() are found in the
main executable. This works because the linker on Mac OS X includes
all symbols of an executable in the global symbol table by default.
To use the same code path on Linux, the application would need to
be built with -Wl,--export-dynamic. To avoid this restriction, we
continue to use weak symbols on Linux systems as before.
Finally this patch extends the existing test to cover all possible
ways of initializing the tool as described by the standard. It
also fixes ompt_finalize() to not call omp_get_thread_num() when
the library is shut down which resulted in hangs on Darwin.
The changes have been tested on Linux to make sure that it passes
the current tests as well as the newly extended one.
Differential Revision: https://reviews.llvm.org/D39801
llvm-svn: 317980
In standalone build, plugins where previously built in their
subdirectory in plugins/ and tests couldn't find them.
Differential Revision: https://reviews.llvm.org/D39920
llvm-svn: 317979
For up-to-date compilers, this assertion is reasonable, but it breaks
compatibility with the typical compiler installed on most systems.
This patch changes the default value to what we had when there was no
compiler support. A warning about the outdated compiler is printed during
runtime, when this point is reached.
Differential Revision: https://reviews.llvm.org/D39890
llvm-svn: 317928
In these places the const attribute seems correct and doesn't
need any other change, so let's do it.
Differential Revision: https://reviews.llvm.org/D39756
llvm-svn: 317798
Allocated memory is typically not 'const' if it needs to be freed.
This patch removes around 50 wrong const attributes, modifies the
corresponding functions and finally gets rid of some const_casts.
These have especially been strange for __kmp_str_fname_free() that
added a 'const' to call __kmp_str_free() which removed it again.
Two minor cleanups that I performed in this process:
* __kmp_tool_libraries now lives in kmp_settings.cpp as it is
used nowhere else.
* __kmp_msg_empty was removed as it was never used and Clang
now complained that it was assigned a string literal that
is 'const char *'.
Differential Revision: https://reviews.llvm.org/D39755
llvm-svn: 317797
If a parallel region is cancelled, execution resumes at the end
of the structured block. That is why this test cannot use the
"normal" macros that print right after inserting the label.
Instead it previously printed the addresses before the pragma
and swapped the checks compared to the other tests.
However, this does not work because FileChecks '*' is greedy
so that RETURN_ADDRESS always matched the second address. This
makes the test fail when an "overflow" occurrs and the first
address matches the value of codeptr_ra.
I discovered this on my MacBook but I'm unable to reproduce the
failure with the current version. Nevertheless we should fix this
problem to avoid that this test fails later after an unrelated change.
Differential Revision: https://reviews.llvm.org/D39708
llvm-svn: 317787
Return addresses are determined based on the address of a label
that is inserted directly after a pragma / API call. In some cases
the tests can assume a known number of instructions between the
addresses. However, the instructions and their encoded lengths
depend on the target that the test is compiled on.
Firstly, this patch refactors the macro print_current_address() to
allow such target dependent modifications and adds information for
the observed instructions on POWER. Secondly, it adapts the related
macro print_fuzzy_address() to reuse much of "hacky" code and fixes
the used formatting strings in the printf() call. Finally, it also
adds documentation about how these macros are intended to work.
Differential Revision: https://reviews.llvm.org/D39699
llvm-svn: 317786
1) Get rid of xaliasify, xexpand and xversionify for KMP_EXPAND_NAME and
KMP_VERSION_SYMBOL. KMP_VERSION_SYMBOL is a combination of xaliasify and
xversionify.
2) Put all attribute and __declspec definitions in kmp_os.h
Differential Revision: https://reviews.llvm.org/D39516
llvm-svn: 317636
The TR6 document is expected to be publically released around November 15.
This patch does not implement OMPT for libomptarget.
Patch by Simon Convent and Joachim Protze
Differential Revision: https://reviews.llvm.org/D39182
llvm-svn: 317436
This is part of the renaming of data types from OpenMP TR4 to TR6
Patch by Simon Convent
Differential Revision: https://reviews.llvm.org/D39326
llvm-svn: 317435
The TR6 document is expected to be publically released around November 15.
This patch does not implement OMPT for libomptarget.
Patch by Simon Convent and Joachim Protze
Differential Revision: https://reviews.llvm.org/D39182
llvm-svn: 317339
This is part of the renaming of data types from OpenMP TR4 to TR6
Patch by Simon Convent
Differential Revision: https://reviews.llvm.org/D39326
llvm-svn: 317338
This was never used in the upstream compiler and was responsible
for some problems with reductions in the clang-ykt fork.
Differential Revision: https://reviews.llvm.org/D39553
llvm-svn: 317214
This is a partial fix for bug 34050.
This prevents callers of omp_set_lock (which does not hold __kmp_global_lock)
from ever seeing an uninitialized version of __kmp_i_lock_table.table.
It does not solve a use-after-free race condition if omp_set_lock obtains a
pointer to __kmp_i_lock_table.table before it is updated and then attempts to
dereference afterwards. That race is far less likely and can be handled in a
separate patch.
The unit test usually segfaults on the current trunk revision. It passes with
the patch.
Patch by Adam Azarchs
Differential Revision: https://reviews.llvm.org/D39439
llvm-svn: 317115
The code is tested to work with latest clang, GNU and Intel compiler. The implementation
is optimized for low overhead when no tool is attached shifting the cost to execution with
tool attached.
This patch does not implement OMPT for libomptarget.
Patch by Simon Convent and Joachim Protze
Differential Revision: https://reviews.llvm.org/D38185
llvm-svn: 317085
Replacing call to __kmp_msg(kmp_ms_fatal,...) with __kmp_fatal(...) caused an
issue when incomplete message is displayed in case an error message is followed
by another message (e.g. by a hint messa)ge. This is because __kmp_fatal()
passes incomplete list of arguments to __kmp_msg().
Patch by Olga Malysheva
Differential Revision: https://reviews.llvm.org/D39248
llvm-svn: 316623
The problem is due to the runtime's threadprivate cleanup code which tries to
access data that was already destroyed by one of the root threads.
__kmp_init_gtid is used as a checker here since it is set to false before actual
resource cleanup is done in __kmp_cleanup().
Patch by Hansang Bae
llvm-svn: 316452
If both KMP_HW_SUBSET and KMP_PLACE_THREADS are set and KMP_PLACE_THREADS gets
parsed first, then the current environment variable parser rejects both and
neither get used. This patch uses the rivals mechanism that is used for other
environment variable groups (e.g., KMP_STACKSIZE, GOMP_STACKSIZE, OMP_STACKSIZE).
If both are set, then it tells the user that it is ignoring KMP_PLACE_THREADS in
favor of KMP_HW_SUBSET. The message about deprecating KMP_PLACE_THREADS when it
is set is still printed regardless.
Differential Revision: https://reviews.llvm.org/D38292
llvm-svn: 315091
Removes semicolons after if {} blocks, function definitions, etc.
I was able to apply the large OMPT patch cleanly on top of this one
with no conflicts.
llvm-svn: 314340
Minor code cleanup of Klocwork issues. Fatal messages are given no return
attribute. Define and use KMP_NORETURN to work for multiple C++ versions.
Patch by Terry Wilmarth
Differential Revision: https://reviews.llvm.org/D37275
llvm-svn: 312538
Cleanup code to remove BUILD_TV and unused code bracketed by it.
Patch by Terry Wilmarth
Differential Revision: https://reviews.llvm.org/D36011
llvm-svn: 311114
Disable default debugging dumps for libomptarget and plugins and move dumps
under control of environment variable LIBOMPTARGET_DEBUG=<integer>. Dumps
are enabled when LIBOMPTARGET_DEBUG is set to a positive integer value.
Debugging dumps are available only in debug build; release build does not
support it.
Differential Revision: https://reviews.llvm.org/D33227
llvm-svn: 310841
This change improves the way threads are spread across cores
when OMP_PROC_BIND=spread is set and no unusual affinity masks are in use.
Differential Revision: https://reviews.llvm.org/D36510
llvm-svn: 310670
We use symbol versioning for GNU-compatibility but libgomp has versioned symbols
only in the shared library but not in the static.
Moreover, version symbols in the static library can cause an error at link time.
Patch by Olga Malysheva
Differential Revision: https://reviews.llvm.org/D36225
llvm-svn: 309877
The original locations can be reached without initializing the lock variable
(td_deque_lock), so it is potentially unsafe. It is guaranteed that the lock
is initialized if the deque (td_deque) is not NULL, and lock functions can be
safely called.
Patch by Hansang Bae
Differential Revision: https://reviews.llvm.org/D36017
llvm-svn: 309875
This change adds a new environment variable, KMP_TEAMS_THREAD_LIMIT, which is
used to set a new global variable, __kmp_teams_max_nth, which is checked when
determining the size and quantity of teams that will be created in the teams
construct. Specifically, it is a limit on the total number of threads in a given
teams construct. It differentiates the limits for the teams construct from the
limits for regular parallel regions (KMP_DEVICE_THREAD_LIMIT/__kmp_max_nth and
OMP_THREAD_LIMIT/__kmp_cg_max_nth). When each individual team is formed, it is
still subject to those limits. After the clauses to the teams construct are
parsed and calculated, we check to make sure we are within this limit, and if
not, reduce num_threads per team and/or number of teams, accordingly. The
default value is set to the number of available processors on the system.
Patch by Terry Wilmarth
Differential Revision: https://reviews.llvm.org/D36009
llvm-svn: 309874
This change fixes the implementation of OMP_THREAD_LIMIT. The implementation of
this previously was not restricted to a contention group (but it should be,
according to the spec), and this is fixed here. A field is added to root thread
to store a counter of the threads in the contention group. An extra check is
added when reserving threads for a parallel region that checks this variable and
compares to threadlimit-var, which is implemented as a new global variable,
kmp_cg_max_nth. Associated settings changes were also made, and clean up of
comments that referred to OMP_THREAD_LIMIT, but should refer to the new
KMP_DEVICE_THREAD_LIMIT (added in an earlier patch).
Patch by Terry Wilmarth
Differential Revision: https://reviews.llvm.org/D35912
llvm-svn: 309319
This change drops in KMP_DEVICE_THREAD_LIMIT to replace KMP_MAX_THREADS. It's
possible there will eventually be a OMP_DEVICE_THREAD_LIMIT, and we need
something to distinguish from OMP_THREAD_LIMIT, which is currently implemented
incorrectly (the fix for that will be added soon in a separate patch).
KMP_ALL_THREADS is deprecated here, but we can keep the "all" option on
KMP_DEVICE_THREAD_LIMIT to support that functionality. KMP_DEVICE_THREAD_LIMIT
now has priority over its deprecated rival KMP_ALL_THREADS. I also cleaned up
some comments that incorrectly referred to non-existent kmp_max_threads variable
instead of kmp_max_nth.
I've left the name of where this setting eventually ends up as
__kmp_max_nth, for now.
This change does not change much in the way of functionality. It does NOT change
OMP_THREAD_LIMIT. It's just cleaning up and setting up for that.
Patch by Terry Wilmarth
Differential Revision: https://reviews.llvm.org/D35860
llvm-svn: 309168
Introduce OPENMP_ENABLE_LIBOMPTARGET which defaults to OFF at the moment.
libomptarget is not yet ready for prime time:
- Offloading to NVIDIA GPUs is not completed yet (compiler, device RTL)
- The generic ELF plugin for offloading to the host (meant for testing)
uses a single instance of the OpenMP runtime (libomp). That is why
omp_is_initial_device() returns 1 which makes the tests fail.
Because of these reasons, we want to disable building (and testing!)
for release 5.0.
See https://bugs.llvm.org/show_bug.cgi?id=33859
Differential Revision: https://reviews.llvm.org/D35719
llvm-svn: 309115
Removed unused __kmp_env_* variables. Also clangified other people's code.
Patch by Terry Wilmarth
Differential Revision: https://reviews.llvm.org/D35808
llvm-svn: 309000
Summary:
The kmp_os.h header is defining the `PAGE_SIZE` macro unconditionally,
even while it is only used directly after its definition, for the
Windows implementation of the `KMP_GET_PAGE_SIZE()` macro.
On at least FreeBSD, but likely all other BSDs too, this macro conflicts
with the one defined in system headers, so remove it, since nothing else
uses it. Make all Unixes use `getpagesize()` instead, and use
`GetSystemInfo()` for the Windows case.
Reviewers: jlpeyton, jcownie, emaste, AndreyChurbanov
Reviewed By: AndreyChurbanov
Subscribers: AndreyChurbanov, hfinkel, zturner
Differential Revision: https://reviews.llvm.org/D35072
llvm-svn: 308355
Summary:
Taskloop implementation is extended by using recursive task scheduling.
Envirable KMP_TASKLOOP_MIN_TASKS added as a manual threshold for the user
to switch from recursive to linear tasks scheduling.
Details:
* The calculations for the loop parameters are moved from __kmp_taskloop_linear
upper level
* Initial calculation is done in the __kmpc_taskloop, further range splitting
is done in the __kmp_taskloop_recur.
* Added threshold to switch from recursive to linear tasks scheduling;
* One half of split range is scheduled as an internal task which just moves
sub-range parameters to the stealing thread that continues recursive
scheduling (if number of tasks still enough), the other half is processed
recursively;
* Internal task duplication routine fixed to assign parent task, that was not
needed when all tasks were scheduled by same thread, but is needed now.
Patch by Andrey Churbanov
Differential Revision: https://reviews.llvm.org/D35273
llvm-svn: 308338
The internal details of this setting are not meant to be user visible and only create confusion.
Differential Revision: https://reviews.llvm.org/D35269
llvm-svn: 308189
Changes are: got all atomics to accept volatile pointers that allowed
to simplify many type conversions. Windows specific code fixed correspondingly.
Differential Revision: https://reviews.llvm.org/D35417
llvm-svn: 308164
The first bit is actually the "untied" flag. That is why the condition was
wrong and has to be inverted to set the flag correctly.
Found and initial patch by Simon Convent!
llvm-svn: 307899
Summary:
On Unix, a .S file is normally an assembly source which must be
preprocessed with a C preprocessor, while a .s file is "plain" assembly.
The former is handled by the compiler driver (cc), the latter is
directly passed to the assembler binary (as).
Because z_Linux_asm.s is supposed to be preprocessed, rename it to .S,
so it can be automatically picked up correctly by build systems.
Reviewers: AndreyChurbanov, emaste, jlpeyton
Reviewed By: AndreyChurbanov
Subscribers: mgorny, openmp-commits
Differential Revision: https://reviews.llvm.org/D35171
llvm-svn: 307680
While importing libomp into the FreeBSD base system we encountered
Clang warnings that "'register' storage class specifier is deprecated
and incompatible with C++1z [-Wdeprecated-register]".
Differential Revision: https://reviews.llvm.org/D35124
llvm-svn: 307441
Changes are: replaced C-style casts with cons_cast and reinterpret_cast;
type of several counters changed to signed; type of parameters of 32-bit and
64-bit AND and OR intrinsics changes to unsigned; changed files formatted
using clang-format version 3.8.1.
Differential Revision: https://reviews.llvm.org/D34759
llvm-svn: 307020
I've found it very difficult to get test/parallel/omp_nested.c to pass
consistently across my build environments. The problem is that it creates N^2
threads (it is testing nested parallel regions), and that often exceeds the
thread limits on systems with many cores. We do raise the process limits in
lit, and that often helps, but if running lit with a smaller number of threads
or on a system where we're otherwise resource constrained, this particular test
tends to fail (because the runtime cannot create a sufficient number of
threads).
This seems to work: if the maximum number of threads is more than some small
number, then cap the number of threads used for the parallel region. The choice
of 4 here is somewhat arbitrary.
Differential Revision: https://reviews.llvm.org/D32033
llvm-svn: 306357
Summary: On BSDs, there is no `libdl.so`, and functions like `dlopen`
are implemented in the main C library instead. Use the `CMAKE_DL_LIBS`
variable instead of hardcoding a dependency on the `dl` library.
Reviewers: grokos, joerg, emaste
Reviewed By: emaste
Subscribers: jlpeyton, mgorny, openmp-commits
Differential Revision: https://reviews.llvm.org/D34632
llvm-svn: 306319
Reset affinity to none (false for proc-bind-var) so that threads in the child
processes are not bound tightly, unless the user explicitly sets this in
KMP_AFFINITY/OMP_PROC_BIND, in child processes. This can improve
performance for scripting languages which fork for parallelism like Python's
multiprocessing module.
Differential Revision: https://reviews.llvm.org/D34154
llvm-svn: 305513
If OpenMP is initialized before fork()-ing occurs and affinity is set to
something like compact, then the master thread will be pinned to a single HW
thread/core after initialization. If the master (or any other thread) then
forks N processes, all N processes will then be pinned to that same single HW
thread/core. To reset the affinity for the new child process, the atfork
handler for the child process can call kmp_set_thread_affinity_mask_initial()
to reset its affinity to the initial affinity of the application before it
re-initializes libomp. The parent process will not be affected and still
keeps its affinity setting.
Differential Revision: https://reviews.llvm.org/D34118
llvm-svn: 305306
Fix static initializers to use the proper unlocked value for the poll
field of the tas and futex locks.
Patch by Terry Wilmarth
Differential Revision: https://reviews.llvm.org/D33794
llvm-svn: 304828
Some code was restructured to move it under KMP_DEBUG. The rest is
formatting changes to fix some things broken by clang-format
Patch by Terry Wilmarth
Differential Revision: https://reviews.llvm.org/D33744
llvm-svn: 304438
With these settings, the create_hwloc_map() method was being called causing an
assert(). After some consideration, it was determined that disabling affinity
explicitly should just disable hwloc as well. i.e., KMP_AFFINITY overrides
KMP_TOPOLOGY_METHOD. This lets the user know that the Hwloc mechanism is being
ignored when KMP_AFFINITY=disabled.
Differential Revision: https://reviews.llvm.org/D33208
llvm-svn: 304344
This change checks if the initial affinity mask is equal to exactly one
Windows processor group's affinity mask. If it is, then the code does not
respect the initial affinity mask and uses the entire machine instead.
The reasoning behind this is that, by default, Windows assigns exactly one
processor group as the initial affinity mask even when there are multiple
Windows processor groups available. User's typically want to use the whole
machine, so we ignore this special case and use the entire machine.
If the initial affinity mask is a proper subset of one group, or spans multiple
groups, then the initial affinity mask is respected since we can assume that the
operating system did not assign this initial affinity mask. This change only
affects machines with multiple processor groups
Differential Revision: https://reviews.llvm.org/D33210
llvm-svn: 304343
An assert() was being tripped when KMP_AFFINITY=respect + Multiple Processor
Groups. Let __kmp_affinity_create_proc_group_map() function be able to create
address2os object which contains a single group by deleting restriction that
process affinity mask must span multiple groups.
llvm-svn: 303101
This patch contains the clang-format and cleanup of the entire code base. Some
of clang-formats changes made the code look worse in places. A best effort was
made to resolve the bulk of these problems, but many remain. Most of the
problems were mangling line-breaks and tabbing of comments.
Patch by Terry Wilmarth
Differential Revision: https://reviews.llvm.org/D32659
llvm-svn: 302929
This patch chagnes the plugin interface so that:
1) future plugins can take advantage of systems with shared CPU/device storage
2) instead of using base addresses, target regions are launched by providing target addresseds and base offsets explicitly.
Differential revision: https://reviews.llvm.org/D33028
llvm-svn: 302663
Older Hwloc libraries (< 1.10.0) don't offer the HWLOC_OBJ_NUMANODE nor
HWLOC_OBJ_PACKAGE types. Instead they are named HWLOC_OBJ_NODE and
HWLOC_OBJ_SOCKET instead. This patch just defines the newer names based on
the older names when using an older Hwloc.
Differential Revision: https://reviews.llvm.org/D32496
llvm-svn: 301349
Without this fix cancellation status for parallel, sections and for persists
across construct boundaries.
Differential Revision: https://reviews.llvm.org/D31419
llvm-svn: 299434
This change slightly improves performance of KMP_YIELD_NOW() macro, by using
_rdtsc() intrinsic function if possible.
Patch by Hansang Bae
Differential Revision: https://reviews.llvm.org/D31008
llvm-svn: 298314
Affinity initialization code expects __kmp_affinity_type has the value
affinity_default by default, but the cleanup code does not properly set the
value back to affinity_default. This may introduce some issues when multiple
roots are trying to initialize/uninitialize the runtime successively.
Patch by Hansang Bae
Differential Revision: https://reviews.llvm.org/D31012
llvm-svn: 298313
This change fixes an assertion failure the in case KMP_AFFINITY is set with
'proclist' specified but without 'explicit'
e.g., KMP_AFFINITY=verbose,proclist=[0-31]
Patch by Olga Malysheva
Differential Revision: https://reviews.llvm.org/D30404
llvm-svn: 297480
Summary:
Bionic didn't get a GNU style strerror_r until Android M. Until then
we unconditionally exposed the POSIX one. Expand the check to account
for this.
Reviewers: pirama, AndreyChurbanov, jlpeyton
Reviewed By: jlpeyton
Subscribers: openmp-commits, srhines
Differential Revision: https://reviews.llvm.org/D30056
llvm-svn: 297235
Add build option LIBOMP_OMP_VERSION=50, 5.0 headers, and add the year/month
associated with OpenMP 5.0 in relevant source locations. Also, remove the
deprecated LIBOMP_OMP_VERSION=41 option.
Patch by Olga Malysheva
Differential Revision: https://reviews.llvm.org/D30450
llvm-svn: 297083
This adds AArch64 support to recently added part of the runtime responsible for offloading to target. This piece of code allows offloading-to-self on AArch64 machines.
Differential Revision: https://reviews.llvm.org/D30644
llvm-svn: 297070
This section of code (__kmp_test_then_* functions) is guarded by
(KMP_ARCH_X86 || KMP_ARCH_X86_64) so it does not make sense to have other
architecture guards inside this section. Non-x86 architectures always
use intrinsics (__sync_*)
llvm-svn: 296525
When using -rtlib=libgcc, the fallback implementation of __atomic_*
builtins is provided via libatomic (included in GCC). However, neither
GCC itself nor clang link libatomic implicitly, and it seems that GCC
upstream expects projects to link it explicitly as necessary.
Since compiler-rt provides __atomic_* builtins directly in the main
library, check if they are provided by the default libraries first.
If they are not, check if -latomic is available to provide them
and add explicit -latomic for tests in this case.
This fixes unresolved __atomic_load() references when running openmp
tests on i386 with libgcc backend.
Differential Revision: https://reviews.llvm.org/D30083
llvm-svn: 296183
Add counter to count number of static_steal for loops
Add counter for number of chunks executed per static_steal for loop
Add counter for number of chunks stolen per static_steal for loop
llvm-svn: 295461
Added test kmp_task_reduction_nest.cpp which has an example of
possible compiler codegen.
Differential Revision: https://reviews.llvm.org/D29600
llvm-svn: 295343
Fixed bug due to which a parent struct was deallocated when one of the struct's pointers was being unmapped.
Differential Revision: https://reviews.llvm.org/D29914
llvm-svn: 295231
This change allows the runtime to turn __kmp_yield() on/off repeatedly on Linux.
This feature was removed when disabling monitor thread, but there are
applications that perform better with this feature on.
Patch by Hansang Bae
Differential Revision: https://reviews.llvm.org/D29227
llvm-svn: 295203
Added new ThreadSanitizer annotations to remove false positives with OpenMP reduction.
Cleaned up Tsan annotations header file from unused annotations.
Patch by Simone Atzeni!
Differential Revision: https://reviews.llvm.org/D29202
llvm-svn: 295158
This change allows setting LIBOMPTARGET_LLVM_LIT_EXECUTABLE and
LIBOMPTARGET_FILECHECK_EXECUTABLE as full path. It also honors
OPENMP_LLVM_TOOLS_DIR which is meant as a common configuration
for both libomp and libomptarget.
Maybe this should be done in a common CMake module, but I'm no expert here.
Differential Revision: https://reviews.llvm.org/D29172
llvm-svn: 294284
This is the patch upstreaming the plugins part of libomptarget (CUDA, generic-elf-64).
Differential Revision: https://reviews.llvm.org/D14253
llvm-svn: 293724
glibc < 2.18 is C99 compliant and only provides the format macros in C++ if
__STDC_FORMAT_MACROS is defined. This change fixes the debug build for
GCC 4.8, GCC 6.2 and Clang 3.9.1 that were previously broken on my machine.
It shows no regression for libc++ >= 4.0.0 which has a fix since September:
http://lists.llvm.org/pipermail/cfe-commits/Week-of-Mon-20160926/171659.html
llvm-svn: 293468
Put the duplicated i_maxmin into traits_t by adding new members max_value and
min_value. Put ___kmp_size_type into traits_t by adding member type_size.
Differential Revision: https://reviews.llvm.org/D28847
llvm-svn: 293316
When the monitor thread is used, most threads in the team directly go to
sleep if the copy of bt_intervals/bt_set is not available in the cache,
and this happens at least once per thread in the wait function, making the
overall performance slightly better.
This change tries to mimic this behavior by using the bt_intervals cache,
which simply keeps the blocktime interval in terms of the platform-dependent
ticks or nanoseconds.
Patch by Hansang Bae
Differential Revision: https://reviews.llvm.org/D28906
llvm-svn: 293312
The lock tables were being reallocated if kmp_set_defaults() was called.
In the env_init code it says that the user should be able to switch between
different KMP_CONSISTENCY_CHECK values which is what this change enables.
llvm-svn: 292349
There is no corresponding free() for this expandable array. The logic is
added in __kmp_cleanup() next to the freeing of __kmp_nested_nth.
llvm-svn: 292348
Clang 4.0 trunk warns:
warning: logical not is only applied to the left hand side of this bitwise operator [-Wlogical-not-parentheses]
This points to a potential bug if the code really wants to check if the single
bit is not set: If for example (buf.edx >> 9) = 2 (has any bit set except the
least significant one), 'logical not' will return 0 which stays 0 after the
'bitwise and'.
To do this correctly we first need to evaluate the 'bitwise and'. In that case
it returns 2 & 1 = 0 which after the 'logical not' evaluates to 1.
Differential Revision: https://reviews.llvm.org/D28599
llvm-svn: 291764
Paul Osmialowski pointed out a double free bug in shutdown code. This patch
Moves the freeing of the implicit task to above the freeing of all fast memory
to prevent the double-free issue.
Differential Revision: https://reviews.llvm.org/D26860
llvm-svn: 287551
Have developer timers use partitioning scheme which also required that some
redundant developer timers be removed in favor of the already existing normal
timers. Move per thread stats initialization to just after global thread id
assignment which is as early as possible. Also put all global stats
initialization code in __kmp_stats_init() and all global stats destruction code
in __kmp_stats_fini().
Differential Revision: https://reviews.llvm.org/D26361
llvm-svn: 286892
This set of changes enables the affinity interface (Either the preexisting
native operating system or HWLOC) to be dynamically set at runtime
initialization. The point of this change is that we were seeing performance
degradations when using HWLOC. This allows the user to use the old affinity
mechanisms which on large machines (>64 cores) makes a large difference in
initialization time.
These changes mostly move affinity code under a small class hierarchy:
KMPAffinity
class Mask {}
KMPNativeAffinity : public KMPAffinity
class Mask : public KMPAffinity::Mask
KMPHwlocAffinity
class Mask : public KMPAffinity::Mask
Since all interface functions (for both affinity and the mask implementation)
are virtual, the implementation can be chosen at runtime initialization.
Differential Revision: https://reviews.llvm.org/D26356
llvm-svn: 286890
This patch allows ThreadSanitizer (Tsan) to verify OpenMP programs.
It means that no false positive will be reported by Tsan when
verifying an OpenMP programs.
This patch introduces annotations within the OpenMP runtime module to
provide information about thread synchronization to the Tsan runtime.
In order to enable the Tsan support when building the runtime, you must
enable the TSAN_SUPPORT option with the following environment variable:
-DLIBOMP_TSAN_SUPPORT=TRUE
The annotations will be enabled in the main shared library
(same mechanism of OMPT).
Patch by Simone Atzeni and Joachim Protze!
Differential Revision: https://reviews.llvm.org/D13072
llvm-svn: 286115
Check Task Scheduling Constraint (TSC) on stealing of untied task.
This is needed because the untied task can produce tied children
those can break TSC if untied is not a descendant of current task.
This can cause live lock on complex tyasking tests
(e.g. kastors/strassen-task-dep).
Differential Revision: https://reviews.llvm.org/D26182
llvm-svn: 285703
Summary:
If directives are used in a macro, clang complains with:
```
src/projects/openmp/runtime/src/kmp_runtime.c:7486:2: error: embedding a directive within macro arguments has undefined behavior [-Werror,-Wembedded-directive]
#if KMP_USE_MONITOR
```
This patch fixes two occurrences of the issue in `kmp_runtime.cpp`.
Reviewers: tlwilmar, jlpeyton, AndreyChurbanov, Hahnfeld
Subscribers: Hahnfeld, openmp-commits
Differential Revision: https://reviews.llvm.org/D25823
llvm-svn: 284728
Function strerror_r() has different signatures in different
implementations of libc: glibc's version returns a char*, while BSDs
and musl return a int. libomp unconditionally assumes glibc on Linux
and thus fails to compile against musl-libc. This patch addresses this
issue.
Differential Revision: https://reviews.llvm.org/D25071
llvm-svn: 284492
New mixed type atomic routines added for regular capture operations as well as
reverse update/capture operations. LHS - all integer and float types (no
complex so far), RHS - float16.
Patch by Olga Malysheva
Differential Revision: https://reviews.llvm.org/D25275
llvm-svn: 284489
This change removes/disables unnecessary code when monitor thread is not used.
Patch by Hansang Bae
Differential Revision: https://reviews.llvm.org/D25102
llvm-svn: 283577
Support finding lit as plain 'lit', which is the name used by setup.py
in LLVM's utils/lit.
Differential Revision: https://reviews.llvm.org/D25072
llvm-svn: 282876
Add check for "45" version to use "201511" string for OpenMP 4.5,
otherwise "200505" is used in Fortran module. Also, fix kmp_openmp_version
variable (used for the debugger, e.g.) and kmp_version_omp_api that is used
in KMP_VERSION=1 output.
Patch by Olga Malysheva
Differential Revision: https://reviews.llvm.org/D24761
llvm-svn: 282868
New routines should be used for atomics like "<int>OP=<float>" when <int> is
unsigned. Using functions __kmpc_atomic_fixed<bits>_<op>_fp) produces incorrect
results
Differential Revision: https://reviews.llvm.org/D24756
llvm-svn: 282509
This change set disables creation of the monitor thread by default. The global
counter maintained by the monitor thread was replaced by logic that uses system
time directly, and cyclic yielding on Linux target was also removed since there
was no clear benefit of using it. Turning on KMP_USE_MONITOR variable (=1)
enables creation of monitor thread again if it is really necessary for some
reasons.
Differential Revision: https://reviews.llvm.org/D24739
llvm-svn: 282507
Fix lit search to correctly respect LIBOMP_LLVM_LIT_EXECUTABLE as full
program path.
The variable passed to find_program() is created by CMake as a cache
variable, and therefore can be directly overriden by the user. Since
this was the design of LIBOMP_LLVM_LIT_EXECUTABLE (as can be deduced
from the error messages) and there is no other use of LIT_EXECUTABLE,
remove the redundant variable and pass LIBOMP_LLVM_LIT_EXECUTABLE
directly to find_program().
Furthermore, the previous code did not work since the HINTS argument
specifies more search directories rather than expected full path.
Quoting the CMake documentation:
> 3. Search the paths specified by the HINTS option. These should be
> paths computed by system introspection, such as a hint provided by
> the location of another item already found. Hard-coded guesses should
> be specified with the PATHS option.
Differential Revision: https://reviews.llvm.org/D24710
llvm-svn: 281887
Introduce a new LIBOMP_INSTALL_VARIABLES cache variable that can be used
to disable creating libgomp and libiomp5 aliases on 'make install'.
Those aliases are undesired e.g. on Gentoo systems where libomp is used
purely by clang.
Differential Revision: https://reviews.llvm.org/D24563
llvm-svn: 281512
Previous differencials D23305-D23310 changed task frame information management only for the kmp interface, but not for the whole gomp interface. This broke some testcases when building with gcc.
This patch fixes the broken task frame information for the gomp interface.
Patch by Joachim Protze!
Differential Revision: https://reviews.llvm.org/D24502
llvm-svn: 281468
In case, the current team is a serialized team (lwt), the frame information should be written to this data structure.
Before, nested serialized teams would overwrite the same task information.
Patch by Joachim Protze!
Differential Revision: https://reviews.llvm.org/D23310
llvm-svn: 281467
The comment already states, that this function should work similarly as __ompt_get_taskinfo.
The function only looked for lwt entries of the current team, but not when unrolling the parents. This fix aligns the implementation to __ompt_get_taskinfo.
The new test case creates a single theaded team (->lwt) and then a nested active team.
Before the innermost print_id(1) would deliver a different team then the outer print_id(0).
Patch by Joachim Protze!
Differential Revision: https://reviews.llvm.org/D23309
llvm-svn: 281466
The exit address is set when execution of a task is started and should be reset as soon as the execution is finished.
Especially for the asm implementation of __kmp_invoke_microtask, resetting in this call would be painfull, so reset just after the invokation.
The testcase shows the effect of this patch:
Before, the implicit barriers at the end of an implicit task would see an exit address for the implicit task.
This barrier is a task scheduling point. Thus, any explicit task scheduled there would see an exit, but no reenter address for the implicit task.
Patch by Joachim Protze!
Differential Revision: https://reviews.llvm.org/D23307
llvm-svn: 281465
The latest OMPT spec changed the semantic of a tasks reenter frame to be the application frame, that will be entered, when the runtime frame drops.
Before it was the last frame in the runtime. This doesn't work for some gcc execution pathes or even clang generated code for :
Since there is no runtime frame between the executed task and the encountering task.
The test case compares exit and reenter addresses against addresses captured in application code
Patch by Joachim Protze!
Differential Revision: https://reviews.llvm.org/D23305
llvm-svn: 281464
OMPT tests can check for right frame information of tasks:
* parent_task_frame was directly printed as a pointer, but actually points to a struct ompt_frame {void*, void*}
* NULL is printed in the beginning of execution and loaded to FileChecker variable [[NULL]]
* implicit tasks now also print their frame information
* macro to print frame address from application
* print task info for barrier begin
Patch by Joachim Protze!
Differential Revision: https://reviews.llvm.org/D23304
llvm-svn: 281463
Rather than checking KMP_CPU_SETSIZE, which doesn't exist when using Hwloc, we
use the get_max_proc() function which can vary based on the operating system.
For example on Windows with multiple processor groups, it might be the case that
the highest bit possible in the bitmask is not equal to the number of hardware
threads on the machine but something higher than that.
Differential Revision: https://reviews.llvm.org/D24206
llvm-svn: 281245
There is a bug in CMakeLists which causes powerpc64le systems to be recognized as big-endian. This patch fixes the issue.
Differential Revision: https://reviews.llvm.org/D23626
llvm-svn: 281068
Implementation of missing OpenMP 4.0 API functions omp_get_default_device and omp_set_default_device.
Also, added support for the environment variable OMP_DEFAULT_DEVICE.
Differential Revision: https://reviews.llvm.org/D23587
llvm-svn: 281065
When affinity isn't supported, __kmp_affinity_compact doesn't exist. The
problem is that in kmp_affinity.h there is a function which uses it without the
proper KMP_AFFINITY_SUPPORTED guard around it. The compiler was smart enough to
ignore it and the function __kmp_affinity_cmp_Address_child_num which relies on
it, but I think it is cleaner to have it under the proper guard. Since the
function is only used in the kmp_affinity.cpp file and there aren't any plans to
have it elsewhere. I have moved it there.
llvm-svn: 280542
the __kmp_affinity_determine_capable() functions are highly operating system
specific. This change has the functions use the type they expect explicitly.
llvm-svn: 280538
In case atomic reduction method is not available (the compiler can't generate
it) the assertion failure occurred if KMP_FORCE_REDUCTION=atomic was specified.
This change replaces the assertion with a warning and sets the reduction method
to the default one - 'critical'.
Patch by Olga Malysheva
Differential Revision: https://reviews.llvm.org/D23990
llvm-svn: 280519
Summary:
On FreeBSD, linking the misc_bugs/omp_foreign_thread_team_reuse.c test
case fails with:
/usr/local/bin/ld: /tmp/omp_foreign_thread_team_reuse-c5e71b.o: undefined reference to symbol 'pthread_create@@FBSD_1.0'
This is because the program is linked without `-lpthread`. Since the
%libomp-compile-and-run macro does not allow that option to be added to
the compile command line, split it up and add the required `-lpthread`
between %libomp-compile and %libomp-run.
Reviewers: jlpeyton, hfinkel, Hahnfeld
Subscribers: Hahnfeld, emaste, openmp-commits
Differential Revision: https://reviews.llvm.org/D23084
llvm-svn: 278036
Consider the following code:
int dep;
#pragma omp target nowait depend(out: dep)
{
sleep(1);
}
#pragma omp task depend(in: dep)
{
printf("Task with dependency\n");
}
printf("Doing some work...\n");
In its current state the runtime will block on the second task and not
continue execution.
Differential Revision: https://reviews.llvm.org/D23116
llvm-svn: 277992
Consider the following code which may be executed by a serial team:
int dep;
#pragma omp target nowait depend(out: dep)
{
sleep(1);
}
#pragma omp task depend(in: dep)
{
#pragma omp target nowait
{
sleep(1);
}
}
Here the explicit task may not be freed until the nested proxy task has
finished. The current code hasn't considered this and called __kmp_free_task
anyway which triggered an assert because of remaining incomplete children:
KMP_DEBUG_ASSERT( TCR_4(taskdata->td_incomplete_child_tasks) == 0 );
Differential Revision: https://reviews.llvm.org/D23115
llvm-svn: 277991
node->dn.task is only filled after the dependencies are already processed.
This currently leads to unhelpful output from KA_TRACE or even a crash
if one enables KMP_SUPPORT_GRAPH_OUTPUT.
llvm-svn: 277717
Summary:
Android does not have pthread_cancel. Disable KMP_CANCEL_THREADS if
__ANDROID__ is defined.
Subscribers: tberghammer, srhines, openmp-commits, danalbert
Differential Revision: https://reviews.llvm.org/D23029
llvm-svn: 277618
This patch enables balanced affinity on machines that do not have
hardware threads and have cores clustered into packages. In facts,
balacing algorithm could be generalized for any arrangement with
at least two levels of hierarchy (depth > 1).
Differential Revision: https://reviews.llvm.org/D22365
llvm-svn: 277212
Summary:
When compiling the runtime library with clang we get warnings like:
```
error: passing an object that undergoes default argument promotion to 'va_start' has undefined behavior [-Werror,-Wvarargs]
va_start( args, id );
^
note: parameter of type 'kmp_i18n_id_t' (aka 'kmp_i18n_id') is declared here
kmp_i18n_id_t id,
```
My understanding is that the va_start macro only gets the promoted type so it won't know what was the exact type of the argument, which can potentially not work for some targets given that the implementation of the the calling convention could not be done properly.
This patch fixes that by using a built-in type in the function signature.
Reviewers: tlwilmar, jlpeyton, AndreyChurbanov
Subscribers: arpith-jacob, carlo.bertolli, caomhin, openmp-commits
Differential Revision: https://reviews.llvm.org/D22427
llvm-svn: 276428
When linking with libhwloc, the ORDERED EPCC test slows down on big
machines (> 48 cores). Performance analysis showed that a cache thrash
was occurring and this padding helps alleviate the problem.
Also, inside the main spin-wait loop in kmp_wait_release.h, we can eliminate
the references to the global shared variables by instead creating a local
variable, oversubscribed and instead checking that.
Differential Revision: http://reviews.llvm.org/D22093
llvm-svn: 274894
These tests are now modeled after the sections nowait test where threads wait
to be released in the first construct (either for or single) and the last thread
skips the last for/single construct and releases those threads. If the test
fails, then it hangs because an unnecessary barrier is executed in between the
constructs.
llvm-svn: 274641
If update_master_only is set the place list is not completely traversed
and therefore this assertion failed. Make it only trigger if
update_master_only is false.
(was introduced by D20539)
Differential Revision: http://reviews.llvm.org/D21925
llvm-svn: 274482
This change fixes an error in comparing the existing schedule on the team to
the new schedule, in the chunk field. Also added additional checks and used
KMP_CHECK_UPDATE where appropriate.
Patch by Terry Wilmarth.
Differential Revision: http://reviews.llvm.org/D21897
llvm-svn: 274371
EPCC Performance of single is considerably worse than plain barrier.
Adding a read-only check to the code before the atomic compare-and-store
helps considerably.
Patch by Terry Wilmarth.
Differential Revision: http://reviews.llvm.org/D21893
llvm-svn: 274369
This rewrite of the omp_sections_nowait.c test file causes it to hang if the
nowait is not respected. If the nowait isn't respected, the lone thread which
can escape the first sections construct will just sleep at a barrier which
shouldn't exist. All reliance on timers is taken out. For good measure, the test
makes sure that all eight sections are executed as well. The test should take no
longer than a few seconds on any modern machine.
Differential Revision: http://reviews.llvm.org/D21842
llvm-svn: 274151
* Incorrect lock value written in __kmp_test_futex_lock
* Incorrect lock value check in tas/futex lock with USE_LOCK_PROFILE on
Patch by Hansang Bae
llvm-svn: 274053
UNICODE and _UNICODE defintions were added in the LLVM CMake build system.
While on Unices, the UNICODE/_UNICODE macros don't cause problems, on Windows
only ittnotify_static.c should be compiled using -DUNICODE. We are still
looking at a proper fix, but this change sets the build back to exactly what it
was doing before. Also, a comment and TODO were added in the src/CMakeLists.txt
file to help explain.
llvm-svn: 274052
That patch made all LLVM projects build with -DUNICODE. However, this doesn't
work for the OpenMP runtime.
But just overriding the flag with -UUNICODE breaks compiling ittnotify_static.c,
which for some reason needs to be compiled with -DUNICIODE. Note that compiling
ittnotify.h with -DUNICODE does not work though.
This seems like a mess. This commit fixes it for now, but it would be great
if someone who works on the OpenMP runtime could fix it properly.
llvm-svn: 273898
Kelvin Li