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
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 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
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
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
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
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
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
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 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
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
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
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:
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
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 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
Jonathan Peyton