Provides AMDGCN and NVPTX specific specialization of getGPUWarpSize,
getGPUThreadID, and getGPUNumThreads methods. Adds tests for AMDGCN
codegen for these methods in generic and simd modes. Also changes the
precondition in InitTempAlloca to be slightly more permissive. Useful for
AMDGCN OpenMP codegen where allocas are created with a cast to an
address space.
Reviewed By: ABataev
Differential Revision: https://reviews.llvm.org/D84260
Refactors CGOpenMPRuntimeNVPTX as CGOpenMPRuntimeGPU to make it a
generalization for OpenMP GPU Codegen. Target specific specialized
methods for NVPTX are defined in class CGOpenMPRuntimeNVPTX. This
paves the way for a clean and maintainable extension to more GPU
targets for OpenMP Codegen.
For original author (git blame) list of CGOpenMPRuntimeGPU code,
look in history of CGOpenMPRuntimeNVPTX.cpp and .h, after this commit.
Reviewed By: ABataev
Differential Revision: https://reviews.llvm.org/D83723
This is the second part loosely extracted from D71179 and cleaned up.
This patch provides semantic analysis support for `omp begin/end declare
variant`, mostly as defined in OpenMP technical report 8 (TR8) [0].
The sema handling makes code generation obsolete as we generate "the
right" calls that can just be handled as usual. This handling also
applies to the existing, albeit problematic, `omp declare variant
support`. As a consequence a lot of unneeded code generation and
complexity is removed.
A major purpose of this patch is to provide proper `math.h`/`cmath`
support for OpenMP target offloading. See PR42061, PR42798, PR42799. The
current code was developed with this feature in mind, see [1].
The logic is as follows:
If we have seen a `#pragma omp begin declare variant match(<SELECTOR>)`
but not the corresponding `end declare variant`, and we find a function
definition we will:
1) Create a function declaration for the definition we were about to generate.
2) Create a function definition but with a mangled name (according to
`<SELECTOR>`).
3) Annotate the declaration with the `OMPDeclareVariantAttr`, the same
one used already for `omp declare variant`, using and the mangled
function definition as specialization for the context defined by
`<SELECTOR>`.
When a call is created we inspect it. If the target has an
`OMPDeclareVariantAttr` attribute we try to specialize the call. To this
end, all variants are checked, the best applicable one is picked and a
new call to the specialization is created. The new call is used instead
of the original one to the base function. To keep the AST printing and
tooling possible we utilize the PseudoObjectExpr. The original call is
the syntactic expression, the specialized call is the semantic
expression.
[0] https://www.openmp.org/wp-content/uploads/openmp-TR8.pdf
[1] https://reviews.llvm.org/D61399#change-496lQkg0mhRN
Reviewers: kiranchandramohan, ABataev, RaviNarayanaswamy, gtbercea, grokos, sdmitriev, JonChesterfield, hfinkel, fghanim, aaron.ballman
Subscribers: bollu, guansong, openmp-commits, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D75779
directive.
According to OpenMP 5.0, The atomic_default_mem_order clause specifies the default memory ordering behavior for atomic constructs that must be provided by an implementation. If the default memory ordering is specified as seq_cst, all atomic constructs on which memory-order-clause is not specified behave as if the seq_cst clause appears. If the default memory ordering is specified as relaxed, all atomic constructs on which memory-order-clause is not specified behave as if the relaxed clause appears.
If the default memory ordering is specified as acq_rel, atomic constructs on which memory-order-clause is not specified behave as if the release clause appears if the atomic write or atomic update operation is specified, as if the acquire clause appears if the atomic read operation is specified, and as if the acq_rel clause appears if the atomic captured update operation is specified.
This removes the OpenMPProcBindClauseKind enum in favor of
llvm::omp::ProcBindKind which lives in OpenMPConstants.h and was
introduced in D70109.
No change in behavior is expected.
Reviewed By: JonChesterfield
Differential Revision: https://reviews.llvm.org/D70289
Summary:
This patch adds support for the registration of the requires directives with the runtime.
Each requires directive clause will enable a particular flag to be set.
The set of flags is passed to the runtime to be checked for compatibility with other such flags coming from other object files.
The registration function is called whenever OpenMP is present even if a requires directive is not present. This helps detect cases in which requires directives are used inconsistently.
Reviewers: ABataev, AlexEichenberger, caomhin
Reviewed By: ABataev, AlexEichenberger
Subscribers: jholewinski, guansong, jfb, jdoerfert, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D60568
llvm-svn: 361298
For the global variables the allocate directive must specify only the
predefined allocator. This allocator must be translated into the correct
form of the address space for the targets that support different address
spaces.
llvm-svn: 356702
memory.
If the variable with the constant non-scalar type is firstprivatized in
the target region, the local copy is created with the data copying.
Instead, we allocate the copy in the constant memory and avoid extra
copying in the outlined target regions. This global copy is used in the
target regions without loss of the performance.
llvm-svn: 355418
A faster way to reduce the values in teams reductions was found, the
codegen is updated to use this faster algorithm and new runtime functions.
llvm-svn: 354479
Emit{Nounwind,}RuntimeCall{,OrInvoke} have been modified to take a
FunctionCallee as an argument, and CreateRuntimeFunction has been
modified to return a FunctionCallee. All callers have been updated.
Additionally, CreateBuiltinFunction is removed, as it was redundant
with CreateRuntimeFunction after some previous changes.
Differential Revision: https://reviews.llvm.org/D57668
llvm-svn: 353184
to reflect the new license.
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: 351636
nvvm_barrier0.
Use runtime functions instead of the direct call to the nvvm intrinsics.
It allows to prevent some dangerous LLVM optimizations, that breaks the
code for the NVPTX target.
llvm-svn: 350328
__kmpc_barrier runtime functions must be marked as convergent to prevent
some dangerous optimizations. Also, for NVPTX target all barriers must
be emitted as simple barriers.
llvm-svn: 348271
modes.
If the region is inside target|teams|distribute region, we can emit the
locations with the correct info for execution mode and runtime mode.
Patch adds this ability to the NVPTX codegen to help the optimizer to
produce better code.
llvm-svn: 347583
For the NVPTX target default locations should be emitted as constants +
additional info must be emitted in the reserved_2 field of the ident_t
structure. The 1st bit controls the execution mode and the 2nd bit
controls use of the lightweight runtime. The combination of the bits for
Non-SPMD mode + lightweight runtime represents special undefined mode,
used outside of the target regions for orphaned directives or functions.
Should allow and additional optimization inside of the target regions.
llvm-svn: 347425
reductions.
Fixed previously committed code for the reduction support in
teams/parallel constructs taking into account new design of the NVPTX
support in the compiler. Teams reduction are not fully functional yet,
it is going to be fixed in the following patches.
llvm-svn: 347081
target|teams|distribute variables.
If the total size of the variables, declared in target|teams|distribute
regions, is less than the maximal size of shared memory available, the
buffer is allocated in the shared memory.
llvm-svn: 346507
target/teams/distribute regions.
Target/teams/distribute regions exist for all the time the kernel is
executed. Thus, if the variable is declared in their context and then
escape it, we can allocate global memory statically instead of
allocating it dynamically.
Patch captures all the globalized variables in target/teams/distribute
contexts, merges them into the records, one per each target region.
Those records are then joined into the union, one per compilation unit
(to save the global memory). Those units are organized into
2 x dimensional arrays, where the first dimension is
the number of blocks per SM and the second one is the number of SMs.
Runtime functions manage this global memory space between the executing
teams.
llvm-svn: 345978
Added support for mapping of lambdas in the target regions. It scans all
the captures by reference in the lambda, implicitly maps those variables
in the target region and then later reinstate the addresses of
references in lambda to the correct addresses of the captured|privatized
variables.
llvm-svn: 345609
Summary: This patch adds a new code generation path for bound sharing directives containing distribute parallel for. The new code generation scheme applies to chunked schedules on distribute and parallel for directives. The scheme simplifies the code that is being generated by eliminating the need for an outer for loop over chunks for both distribute and parallel for directives. In the case of distribute it applies to any sized chunk while in the parallel for case it only applies when chunk size is 1.
Reviewers: ABataev, caomhin
Reviewed By: ABataev
Subscribers: jholewinski, guansong, cfe-commits
Differential Revision: https://reviews.llvm.org/D53448
llvm-svn: 345509
if the function has globalized variables and called in context of
target/teams/distribute regions, it does not need to globalize 32
copies of the same variables for memory coalescing, it is enough to
have just one copy, because there is parallel region.
Patch does this by adding call for `__kmpc_parallel_level` function and
checking its return value. If the code sees that the parallel level is
0, then only one variable is allocated, not 32.
llvm-svn: 344356
Added support for memory coalescing for better performance for
globalized variables. From now on all the globalized variables are
represented as arrays of 32 elements and each thread accesses these
elements using `tid & 31` as index.
llvm-svn: 344049
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 target'.
All the functions, referenced in implicit|explicit target regions must
be emitted during code emission for the device.
llvm-svn: 341093
If the target construct can be executed in SPMD mode + it is a loop
based directive with static scheduling, we can use lightweight runtime
support.
llvm-svn: 340953
parallel region.
If the current construct requires sharing of the local variable in the
inner parallel region, this variable must be globalized to avoid
runtime crash.
llvm-svn: 335285
Added initial support for L2 parallelism in SPMD mode. Note, though,
that the orphaned parallel directives are not currently supported in
SPMD mode.
llvm-svn: 332016
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
Added initial codegen for level 2, 3 etc. parallelism. Currently, all
the second, the third etc. parallel regions will run sequentially.
llvm-svn: 331642
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 generic codegen is enabled and private copy of the original
variable escapes the declaration context, this private copy should be
globalized just like it was the original variable.
llvm-svn: 327985
If the variable is captured by value and the corresponding parameter in
the outlined function escapes its declaration context, this parameter
must be globalized. To globalize it we need to get the address of the
original parameter, load the value, store it to the global address and
use this global address instead of the original.
Patch improves globalization for parallel|teams regions + functions in
declare target regions.
llvm-svn: 327654
Summary: Remove this scheme for now since it will be covered by another more generic scheme using global memory. This code will be worked into an optimization for the generic data sharing scheme. Removing this completely and then adding it via future patches will make all future data sharing patches cleaner.
Reviewers: ABataev, carlo.bertolli, caomhin
Reviewed By: ABataev
Subscribers: jholewinski, guansong, cfe-commits
Differential Revision: https://reviews.llvm.org/D43625
llvm-svn: 326948
Summary:
This patch is part of the development effort to add support in the current OpenMP GPU offloading implementation for implicitly sharing variables between a target region executed by the team master thread and the worker threads within that team.
This patch is the first of three required for successfully performing the implicit sharing of master thread variables with the worker threads within a team. The remaining two patches are:
- Patch D38978 to the LLVM NVPTX backend which ensures the lowering of shared variables to an device memory which allows the sharing of references;
- Patch (coming soon) is a patch to libomptarget runtime library which ensures that a list of references to shared variables is properly maintained.
A simple code snippet which illustrates an implicit data sharing situation is as follows:
```
#pragma omp target
{
// master thread only
int v;
#pragma omp parallel
{
// worker threads
// use v
}
}
```
Variable v is implicitly shared from the team master thread which executes the code in between the target and parallel directives. The worker threads must operate on the latest version of v, including any updates performed by the master.
The code generated in this patch relies on the LLVM NVPTX patch (mentioned above) which prevents v from being lowered in the thread local memory of the master thread thus making the reference to this variable un-shareable with the workers. This ensures that the code generated by this patch is correct.
Since the parallel region is outlined the passing of arguments to the outlined regions must preserve the original order of arguments. The runtime therefore maintains a list of references to shared variables thus ensuring their passing in the correct order. The passing of arguments to the outlined parallel function is performed in a separate function which the data sharing infrastructure constructs in this patch. The function is inlined when optimizations are enabled.
Reviewers: hfinkel, carlo.bertolli, arpith-jacob, Hahnfeld, ABataev, caomhin
Reviewed By: ABataev
Subscribers: cfe-commits, jholewinski
Differential Revision: https://reviews.llvm.org/D38976
llvm-svn: 318773
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
Saiyedul Islam