Summary: Set default schedule for parallel for loops to schedule(static, 1) when using SPMD mode on the NVPTX device offloading toolchain to ensure coalescing.
Reviewers: ABataev, Hahnfeld, caomhin
Reviewed By: ABataev
Subscribers: jholewinski, guansong, cfe-commits
Differential Revision: https://reviews.llvm.org/D52629
llvm-svn: 343260
Summary: For the OpenMP NVPTX toolchain choose a default distribute schedule that ensures coalescing on the GPU when in SPMD mode. This significantly increases the performance of offloaded target code and reduces the number of registers used on the GPU side.
Reviewers: ABataev, caomhin, Hahnfeld
Reviewed By: ABataev, Hahnfeld
Subscribers: Hahnfeld, jholewinski, guansong, cfe-commits
Differential Revision: https://reviews.llvm.org/D52434
llvm-svn: 343253
declare reduction.
If the declare reduction construct with the non-dependent type is
defined in the template construct, the compiler might crash on the
template instantition. Reworked the whole instantiation scheme for the
declare reduction constructs to fix this problem correctly.
llvm-svn: 342151
declare target.
According to OpenMP 5.0, variables captured in lambdas in declare target
regions must be considered as implicitly declare target.
llvm-svn: 339152
If the declare target link entries are created but not used, the
compiler will produce an error message. Patch improves handling of such
situations + improves checks for possibly lost declare target variables.
llvm-svn: 337207
This is similar to the LLVM change https://reviews.llvm.org/D46290.
We've been running doxygen with the autobrief option for a couple of
years now. This makes the \brief markers into our comments
redundant. Since they are a visual distraction and we don't want to
encourage more \brief markers in new code either, this patch removes
them all.
Patch produced by
for i in $(git grep -l '\@brief'); do perl -pi -e 's/\@brief //g' $i & done
for i in $(git grep -l '\\brief'); do perl -pi -e 's/\\brief //g' $i & done
Differential Revision: https://reviews.llvm.org/D46320
llvm-svn: 331834
Some symbols are not allowed to be used as names on some targets. Patch
ries to unify the emission of the names of LLVM globals so they could be
used on different targets.
llvm-svn: 331358
Found via codespell -q 3 -I ../clang-whitelist.txt
Where whitelist consists of:
archtype
cas
classs
checkk
compres
definit
frome
iff
inteval
ith
lod
methode
nd
optin
ot
pres
statics
te
thru
Patch by luzpaz! (This is a subset of D44188 that applies cleanly with a few
files that have dubious fixes reverted.)
Differential revision: https://reviews.llvm.org/D44188
llvm-svn: 329399
variables.
Added emission of the offloading data sections for the variables within
declare target regions + fixes emission of the declare target variables
marked as declare target not within the declare target region.
llvm-svn: 328888
When the declare target variables are emitted for the device,
constructors|destructors for these variables must emitted and registered
by the runtime in the offloading sections.
llvm-svn: 328705
If the link clause is used on the declare target directive, the object
should be linked on target or target data directives, not during the
codegen. Patch adds support for this clause.
llvm-svn: 328544
Added initial codegen for device side of declarations inside `omp
declare target` construct + codegen for implicit `declare target`
functions, which are used in the target regions.
llvm-svn: 327636
variables.
If the task has reduction construct and this construct for some variable
requires unique threadprivate storage, we may generate different names
for variables used in taskgroup task_reduction clause and in task
in_reduction clause. Patch fixes this problem.
llvm-svn: 326827
only.
Added support for -fopenmp-simd option that allows compilation of
simd-based constructs without emission of OpenMP runtime calls.
llvm-svn: 321560
If both taskloop and task directives are used at the same time in one
program, we may ran into the situation when the particular type for task
directive is reused for taskloop directives. Patch fixes this problem.
llvm-svn: 315464
__kmpc_for_static_fini().
Added special flags for calls of __kmpc_for_static_fini(), like previous
ly for __kmpc_for_static_init(). Added flag OMP_IDENT_WORK_DISTRIBUTE
for distribute cnstruct, OMP_IDENT_WORK_SECTIONS for sections-based
constructs and OMP_IDENT_WORK_LOOP for loop-based constructs in
location flags.
llvm-svn: 312642
__kmpc_for_static_init().
OpenMP 5.0 will include OpenMP Tools interface that requires distinguishing different worksharing constructs.
Since the same entry point (__kmp_for_static_init(ident_t *loc,
kmp_int32 global_tid,........)) is called in case static
loop/sections/distribute it is suggested using 'flags' field of the
ident_t structure to pass the type of the construct.
llvm-svn: 310865
Arguments, passed to the outlined function, must have correct address
space info for proper Debug info support. Patch sets global address
space for arguments that are mapped and passed by reference.
Also, cuda-gdb does not handle reference types correctly, so reference
arguments are represented as pointers.
llvm-svn: 310387
Arguments, passed to the outlined function, must have correct address
space info for proper Debug info support. Patch sets global address
space for arguments that are mapped and passed by reference.
Also, cuda-gdb does not handle reference types correctly, so reference
arguments are represented as pointers.
llvm-svn: 310377
Arguments, passed to the outlined function, must have correct address
space info for proper Debug info support. Patch sets global address
space for arguments that are mapped and passed by reference.
Also, cuda-gdb does not handle reference types correctly, so reference
arguments are represented as pointers.
llvm-svn: 310360
Arguments, passed to the outlined function, must have correct address
space info for proper Debug info support. Patch sets global address
space for arguments that are mapped and passed by reference.
Also, cuda-gdb does not handle reference types correctly, so reference
arguments are represented as pointers.
llvm-svn: 310104
[OpenMP] Initial implementation of code generation for pragma 'distribute parallel for' on host
https://reviews.llvm.org/D29508
This patch makes the following additions:
It abstracts away loop bound generation code from procedures associated with pragma 'for' and loops in general, in such a way that the same procedures can be used for 'distribute parallel for' without the need for a full re-implementation.
It implements code generation for 'distribute parallel for' and adds regression tests. It includes tests for clauses.
It is important to notice that most of the clauses are implemented as part of existing procedures. For instance, firstprivate is already implemented for 'distribute' and 'for' as separate pragmas. As the implementation of 'distribute parallel for' is based on the same procedures, then we automatically obtain implementation for such clauses without the need to add new code. However, this requires regression tests that verify correctness of produced code.
llvm-svn: 301340
https://reviews.llvm.org/D29508
This patch makes the following additions:
1. It abstracts away loop bound generation code from procedures associated with pragma 'for' and loops in general, in such a way that the same procedures can be used for 'distribute parallel for' without the need for a full re-implementation.
2. It implements code generation for 'distribute parallel for' and adds regression tests. It includes tests for clauses.
It is important to notice that most of the clauses are implemented as part of existing procedures. For instance, firstprivate is already implemented for 'distribute' and 'for' as separate pragmas. As the implementation of 'distribute parallel for' is based on the same procedures, then we automatically obtain implementation for such clauses without the need to add new code. However, this requires regression tests that verify correctness of produced code.
Looking forward to comments.
llvm-svn: 301223
This patch implements codegen for the reduction clause on
any parallel construct for elementary data types. An efficient
implementation requires hierarchical reduction within a
warp and a threadblock. It is complicated by the fact that
variables declared in the stack of a CUDA thread cannot be
shared with other threads.
The patch creates a struct to hold reduction variables and
a number of helper functions. The OpenMP runtime on the GPU
implements reduction algorithms that uses these helper
functions to perform reductions within a team. Variables are
shared between CUDA threads using shuffle intrinsics.
An implementation of reductions on the NVPTX device is
substantially different to that of CPUs. However, this patch
is written so that there are minimal changes to the rest of
OpenMP codegen.
The implemented design allows the compiler and runtime to be
decoupled, i.e., the runtime does not need to know of the
reduction operation(s), the type of the reduction variable(s),
or the number of reductions. The design also allows reuse of
host codegen, with appropriate specialization for the NVPTX
device.
While the patch does introduce a number of abstractions, the
expected use case calls for inlining of the GPU OpenMP runtime.
After inlining and optimizations in LLVM, these abstractions
are unwound and performance of OpenMP reductions is comparable
to CUDA-canonical code.
Patch by Tian Jin in collaboration with Arpith Jacob
Reviewers: ABataev
Differential Revision: https://reviews.llvm.org/D29758
llvm-svn: 295333
This patch implements codegen for the reduction clause on
any parallel construct for elementary data types. An efficient
implementation requires hierarchical reduction within a
warp and a threadblock. It is complicated by the fact that
variables declared in the stack of a CUDA thread cannot be
shared with other threads.
The patch creates a struct to hold reduction variables and
a number of helper functions. The OpenMP runtime on the GPU
implements reduction algorithms that uses these helper
functions to perform reductions within a team. Variables are
shared between CUDA threads using shuffle intrinsics.
An implementation of reductions on the NVPTX device is
substantially different to that of CPUs. However, this patch
is written so that there are minimal changes to the rest of
OpenMP codegen.
The implemented design allows the compiler and runtime to be
decoupled, i.e., the runtime does not need to know of the
reduction operation(s), the type of the reduction variable(s),
or the number of reductions. The design also allows reuse of
host codegen, with appropriate specialization for the NVPTX
device.
While the patch does introduce a number of abstractions, the
expected use case calls for inlining of the GPU OpenMP runtime.
After inlining and optimizations in LLVM, these abstractions
are unwound and performance of OpenMP reductions is comparable
to CUDA-canonical code.
Patch by Tian Jin in collaboration with Arpith Jacob
Reviewers: ABataev
Differential Revision: https://reviews.llvm.org/D29758
llvm-svn: 295319
This patch adds support for codegen of 'target parallel' on the host.
It is also the first combined directive that requires two or more
captured statements. Support for this functionality is included in
the patch.
A combined directive such as 'target parallel' has two captured
statements, one for the 'target' and the other for the 'parallel'
region. Two captured statements are required because each has
different implicit parameters (see SemaOpenMP.cpp). For example,
the 'parallel' has 'global_tid' and 'bound_tid' while the 'target'
does not. The patch adds support for handling multiple captured
statements based on the combined directive.
When codegen'ing the 'target parallel' directive, the 'target'
outlined function is created using the outer captured statement
and the 'parallel' outlined function is created using the inner
captured statement.
Reviewers: ABataev
Differential Revision: https://reviews.llvm.org/D28753
llvm-svn: 292419
This patch adds support for codegen of 'target parallel' on the host.
It is also the first combined directive that requires two or more
captured statements. Support for this functionality is included in
the patch.
A combined directive such as 'target parallel' has two captured
statements, one for the 'target' and the other for the 'parallel'
region. Two captured statements are required because each has
different implicit parameters (see SemaOpenMP.cpp). For example,
the 'parallel' has 'global_tid' and 'bound_tid' while the 'target'
does not. The patch adds support for handling multiple captured
statements based on the combined directive.
When codegen'ing the 'target parallel' directive, the 'target'
outlined function is created using the outer captured statement
and the 'parallel' outlined function is created using the inner
captured statement.
Reviewers: ABataev
Differential Revision: https://reviews.llvm.org/D28753
llvm-svn: 292374
Summary:
This patch introduces support for the execution of parallel constructs in a target
region on the NVPTX device. Parallel regions must be in the lexical scope of the
target directive.
The master thread in the master warp signals parallel work for worker threads in worker
warps on encountering a parallel region.
Note: The patch does not yet support capture of arguments in a parallel region so
the test cases are simple.
Reviewers: ABataev
Differential Revision: https://reviews.llvm.org/D28145
llvm-svn: 291565
Summary:
This patch adds two fields to the offload entry descriptor. One field is meant to signal Ctors/Dtors and `link` global variables, and the other is reserved for runtime library use.
Currently, these fields are only filled with zeros in the current code generation, but that will change when `declare target` is added.
The reason, we are adding these fields now is to make the code generation consistent with the runtime library proposal under review in https://reviews.llvm.org/D14031.
Reviewers: ABataev, hfinkel, carlo.bertolli, kkwli0, arpith-jacob, Hahnfeld
Subscribers: cfe-commits, caomhin, jholewinski
Differential Revision: https://reviews.llvm.org/D28298
llvm-svn: 291124
Summary: This patch adds support for the use_device_ptr clause. It includes changes in SEMA that could not be tested without codegen, namely, the use of the first private logic and mappable expressions support.
Reviewers: hfinkel, carlo.bertolli, arpith-jacob, kkwli0, ABataev
Subscribers: caomhin, cfe-commits
Differential Revision: https://reviews.llvm.org/D22691
llvm-svn: 276977
Summary: This patch implements the code generation for the `target update` directive. The implemntation relies on the logic already in place for target data standalone directives, i.e. target enter/exit data.
Reviewers: hfinkel, carlo.bertolli, arpith-jacob, kkwli0, ABataev
Subscribers: caomhin, cfe-commits
Differential Revision: http://reviews.llvm.org/D20650
llvm-svn: 270886
directives.
OpenMP 4.5 supports clause 'priority' in task-based directives. Patch
adds initial codegen support for this clause in codegen.
llvm-svn: 269050
If private variables require destructors call at the deletion of the
task, additional flag in task flags must be set. Patch fixes this
problem.
llvm-svn: 269039
schedule modifiers.
Runtime library expects some additional data in schedule argument for
loop-based directives, that have additional schedule modifiers
'monotonic|nonmonotonic'.
llvm-svn: 269035
OpenMP 4.5 adds taskloop/taskloop simd directives. These directives
allow to use lastprivate clause. Patch adds codegen for this clause.
llvm-svn: 268618
directive.
OpenMP 4.5 defines 'taskloop' directive and 2 additional clauses
'grainsize' and 'num_tasks' for this directive. Patch adds codegen for
these clauses.
These clauses are generated as arguments of the '__kmpc_taskloop'
libcall and are encoded the following way:
void __kmpc_taskloop(ident_t *loc, int gtid, kmp_task_t *task, int if_val, kmp_uint64 *lb, kmp_uint64 *ub, kmp_int64 st, int nogroup, int sched, kmp_uint64 grainsize, void *task_dup);
If 'grainsize' is specified, 'sched' argument must be set to '1' and
'grainsize' argument must be set to the value of the 'grainsize' clause.
If 'num_tasks' is specified, 'sched' argument must be set to '2' and
'grainsize' argument must be set to the value of the 'num_tasks' clause.
It is possible because these 2 clauses are mutually exclusive and can't
be used at the same time on the same directive.
If none of these clauses is specified, 'sched' argument must be set to
'0'.
llvm-svn: 267862
Summary:
This patch adds support for the target exit data directive code generation.
Given that, apart from the employed runtime call, target exit data requires the same code generation pattern as target enter data, the OpenMP codegen entry point was renamed and reused for both.
Reviewers: hfinkel, carlo.bertolli, arpith-jacob, kkwli0, ABataev
Subscribers: cfe-commits, fraggamuffin, caomhin
Differential Revision: http://reviews.llvm.org/D17369
llvm-svn: 267814
Summary: This patch adds support for the target enter data directive code generation.
Reviewers: hfinkel, carlo.bertolli, arpith-jacob, kkwli0, ABataev
Subscribers: cfe-commits, fraggamuffin, caomhin
Differential Revision: http://reviews.llvm.org/D17368
llvm-svn: 267812
Summary:
This patch adds support for the target data directive code generation.
Part of the already existent functionality related with data maps is moved to a new function so that it could be reused.
Reviewers: hfinkel, carlo.bertolli, arpith-jacob, kkwli0, ABataev
Subscribers: cfe-commits, fraggamuffin, caomhin
Differential Revision: http://reviews.llvm.org/D17367
llvm-svn: 267811
The taskloop construct specifies that the iterations of one or more associated loops will be executed in parallel using OpenMP tasks. The iterations are distributed across tasks created by the construct and scheduled to be executed.
The next code will be generated for the taskloop directive:
#pragma omp taskloop num_tasks(N) lastprivate(j)
for( i=0; i<N*GRAIN*STRIDE-1; i+=STRIDE ) {
int th = omp_get_thread_num();
#pragma omp atomic
counter++;
#pragma omp atomic
th_counter[th]++;
j = i;
}
Generated code:
task = __kmpc_omp_task_alloc(NULL,gtid,1,sizeof(struct
task),sizeof(struct shar),&task_entry);
psh = task->shareds;
psh->pth_counter = &th_counter;
psh->pcounter = &counter;
psh->pj = &j;
task->lb = 0;
task->ub = N*GRAIN*STRIDE-2;
task->st = STRIDE;
__kmpc_taskloop(
NULL, // location
gtid, // gtid
task, // task structure
1, // if clause value
&task->lb, // lower bound
&task->ub, // upper bound
STRIDE, // loop increment
0, // 1 if nogroup specified
2, // schedule type: 0-none, 1-grainsize, 2-num_tasks
N, // schedule value (ignored for type 0)
(void*)&__task_dup_entry // tasks duplication routine
);
llvm-svn: 267395
If the untied clause is present on a task construct, any thread in the
team can resume the task region after a suspension. Patch adds proper
codegen for untied tasks.
llvm-svn: 266853
If the untied clause is present on a task construct, any thread in the
team can resume the task region after a suspension. Patch adds proper
codegen for untied tasks.
llvm-svn: 266754
If the untied clause is present on a task construct, any thread in the team can resume the task region after a suspension. Patch adds proper codegen for untied tasks.
llvm-svn: 266722
This patch implements the teams directive for the NVPTX backend. It is different from the host code generation path as it:
Does not call kmpc_fork_teams. All necessary teams and threads are started upon touching the target region, when launching a CUDA kernel, and their execution is coordinated through sequential and parallel regions within the target region.
Does not call kmpc_push_num_teams even if a num_teams of thread_limit clause is present. Setting the number of teams and the thread limit is implemented by the nvptx-related runtime.
Please note that I am now passing a Clang Expr * to emitPushNumTeams instead of the originally chosen llvm::Value * type. The reason for that is that I want to avoid emitting expressions for num_teams and thread_limit if they are not needed in the target region.
http://reviews.llvm.org/D17963
llvm-svn: 265304
Solution unifies interface of RegionCodeGenTy type to allow insert
runtime-specific code before/after main codegen action defined in
CGStmtOpenMP.cpp file. Runtime should not define its own RegionCodeGenTy
for general OpenMP directives, but must be allowed to insert its own
(required) code to support target specific codegen.
llvm-svn: 264700
Solution unifies interface of RegionCodeGenTy type to allow insert
runtime-specific code before/after main codegen action defined in
CGStmtOpenMP.cpp file. Runtime should not define its own RegionCodeGenTy
for general OpenMP directives, but must be allowed to insert its own
(required) code to support target specific codegen.
llvm-svn: 264576
Solution unifies interface of RegionCodeGenTy type to allow insert
runtime-specific code before/after main codegen action defined in
CGStmtOpenMP.cpp file. Runtime should not define its own RegionCodeGenTy
for general OpenMP directives, but must be allowed to insert its own
(required) code to support target specific codegen.
llvm-svn: 264569
Summary:
This patch adds base support for codegen of the target directive on the NVPTX device.
Reviewers: ABataev
Differential Revision: http://reviews.llvm.org/D17877
Reworked test case after buildbot failure on windows.
Updated patch to integrate r263837 and test case nvptx_target_firstprivate_codegen.cpp.
llvm-svn: 264018
Summary:
Reworked test case after buildbot failure on windows.
This patch adds base support for codegen of the target directive on the NVPTX device.
Reviewers: ABataev
Differential Revision: http://reviews.llvm.org/D17877
llvm-svn: 263783
OpenMP 4.0 allows to define custom reduction operations using '#pragma
omp declare reduction' construct. Patch allows to use this custom
defined reduction operations in 'reduction' clauses.
llvm-svn: 263701
Summary:
This patch adds base support for codegen of the target directive on the NVPTX device.
Reviewers: ABataev
Differential Revision: http://reviews.llvm.org/D17877
llvm-svn: 263587
Summary:
This patch adds base support for codegen of the target directive on the NVPTX device.
Reviewers: ABataev
Differential Revision: http://reviews.llvm.org/D17877
llvm-svn: 263552
This patch provide basic implementation of codegen for teams directive, excluding all clauses except dist_schedule. It also fixes parts of AST reader/writer to enable correct pre-compiled header handling.
http://reviews.llvm.org/D17170
llvm-svn: 262832
This patch provide basic implementation of codegen for teams directive, excluding all clauses except dist_schedule. It also fixes parts of AST reader/writer to enable correct pre-compiled header handling.
http://reviews.llvm.org/D17170
llvm-svn: 262741
Summary:
Unlike other outlined regions in OpenMP, offloading entry points have to have be visible (external linkage) for the device side. Using dots in the names of the entries can be therefore problematic for some toolchains, e.g. NVPTX.
Also the patch drops the column information in the unique name of the entry points. The parsing of directives ignore unknown tokens, preventing several target regions to be implemented in the same line. Therefore, the line information is sufficient for the name to be unique. Also, the preprocessor printer does not preserve the column information, causing offloading-entry detection issues if the host uses an integrated preprocessor and the target doesn't (or vice versa).
Reviewers: hfinkel, arpith-jacob, carlo.bertolli, kkwli0, ABataev
Subscribers: cfe-commits, fraggamuffin, caomhin
Differential Revision: http://reviews.llvm.org/D17179
llvm-svn: 260837
This patch attempts to fix the regressions identified when the patch was committed initially.
Thanks to Michael Liao for identifying the fix in the offloading metadata generation
related with side effects in evaluation of function arguments.
llvm-svn: 256933
Gheorghe-Teodor Bercea