[PATCH] D60883: [OpenMP] Avoid emitting maps for target link variables when unified memory is used

gtbercea created this revision. gtbercea added reviewers: ABataev, caomhin. Herald added subscribers: cfe-commits, jdoerfert, guansong. Herald added a project: clang. This patch avoids the emission of maps for target link variables when unified memory is present. Repository: rC Clang https:/

[PATCH] D60886: [OpenMP][NFC] Fix requires target test.

gtbercea created this revision. gtbercea added a reviewer: ABataev. Herald added subscribers: cfe-commits, jdoerfert, guansong. Herald added a project: clang. Fix requires target test. Repository: rC Clang https://reviews.llvm.org/D60886 Files: test/OpenMP/requires_target_messages.cpp In

[PATCH] D60907: [OpenMP][WIP] Add math functions support in OpenMP offloading

gtbercea created this revision. gtbercea added reviewers: ABataev, hfinkel, caomhin. Herald added subscribers: cfe-commits, jdoerfert, guansong, mgorny. Herald added a project: clang. gtbercea added a reviewer: tra. gtbercea added parent revisions: D60906: [OpenMP][libomptarget][WIP] Add math func

[PATCH] D60907: [OpenMP][WIP] Add math functions support in OpenMP offloading

gtbercea updated this revision to Diff 195915. gtbercea edited the summary of this revision. gtbercea added a comment. - Address comments. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D60907/new/ https://reviews.llvm.org/D60907 Files: include/clang/Driver/Tool

[PATCH] D60907: [OpenMP][WIP] Add math functions support in OpenMP offloading

gtbercea added a comment. In D60907#1473406 , @Hahnfeld wrote: > So the scheme is: `pow` is defined in `__clang_openmp_math.h` to call > `__kmpc_pow`. This lives in `libomptarget-nvptx` (both bc and static lib) and > just calls `pow` which works because

[PATCH] D60907: [OpenMP][WIP] Add math functions support in OpenMP offloading

gtbercea added a subscriber: gregrodgers. gtbercea added a comment. @gregrodgers Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D60907/new/ https://reviews.llvm.org/D60907 ___ cfe-commits mailing list cfe-commits@lists

[PATCH] D60907: [OpenMP][WIP] Add math functions support in OpenMP offloading

gtbercea updated this revision to Diff 196619. gtbercea added a comment. - Use macros. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D60907/new/ https://reviews.llvm.org/D60907 Files: include/clang/Driver/ToolChain.h lib/Driver/ToolChains/Clang.cpp lib/Driv

[PATCH] D60907: [OpenMP] Add math functions support in OpenMP offloading

gtbercea added a comment. Ping @hfinkel @tra Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D60907/new/ https://reviews.llvm.org/D60907 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi

[PATCH] D60907: [OpenMP] Add math functions support in OpenMP offloading

gtbercea added a comment. In D60907#1479142 , @hfinkel wrote: > In D60907#1479118 , @gtbercea wrote: > > > Ping @hfinkel @tra > > > The last two comments in D47849 indicated > expl

[PATCH] D60907: [OpenMP] Add math functions support in OpenMP offloading

gtbercea added a comment. In D60907#1479298 , @gtbercea wrote: > In D60907#1479142 , @hfinkel wrote: > > > In D60907#1479118 , @gtbercea > > wrote: > > > > > Ping @hfinkel @

[PATCH] D60907: [OpenMP] Add math functions support in OpenMP offloading

gtbercea added a comment. In D60907#1479142 , @hfinkel wrote: > In D60907#1479118 , @gtbercea wrote: > > > Ping @hfinkel @tra > > > The last two comments in D47849 indicated > expl

[PATCH] D60907: [OpenMP] Add math functions support in OpenMP offloading

gtbercea updated this revision to Diff 196725. gtbercea added a comment. - Update patch. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D60907/new/ https://reviews.llvm.org/D60907 Files: include/clang/Driver/ToolChain.h lib/Driver/ToolChains/Clang.cpp lib/Dr

[PATCH] D60568: [OpenMP] Add support for registering requires directives with the runtime

gtbercea updated this revision to Diff 196945. gtbercea marked 2 inline comments as done. gtbercea added a comment. - Add tests. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D60568/new/ https://reviews.llvm.org/D60568 Files: lib/CodeGen/CGOpenMPRuntime.cpp l

[PATCH] D61399: [OpenMP][Clang] Support for target math functions

gtbercea created this revision. gtbercea added reviewers: hfinkel, caomhin, ABataev, tra. Herald added subscribers: cfe-commits, guansong, mgorny. Herald added a project: clang. In this patch we propose a temporary solution to resolving math functions for the NVPTX toolchain, temporary until Open

[PATCH] D61399: [OpenMP][Clang] Support for target math functions

gtbercea updated this revision to Diff 197638. gtbercea edited the summary of this revision. gtbercea added a comment. - Minor fixes. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D61399/new/ https://reviews.llvm.org/D61399 Files: lib/Driver/ToolChains/Clang.cp

[PATCH] D61399: [OpenMP][Clang] Support for target math functions

gtbercea updated this revision to Diff 197798. gtbercea added a comment. - Clean-up. Add header. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D61399/new/ https://reviews.llvm.org/D61399 Files: lib/Driver/ToolChains/Clang.cpp lib/Headers/CMakeLists.txt lib/

[PATCH] D61474: [CUDA][Clang][Bugfix] Add missing CUDA 9.2 case

gtbercea created this revision. gtbercea added reviewers: tra, ABataev, caomhin. Herald added subscribers: cfe-commits, jdoerfert. Herald added a project: clang. The bug was reported on the OpenMP-dev list: .../obj-release/lib/clang/9.0.0/include/__clang_cuda_intrinsics.h:173:35: error: '__nvvm_

[PATCH] D61399: [OpenMP][Clang] Support for target math functions

gtbercea updated this revision to Diff 197911. gtbercea added a comment. - Address comments. - Add math and cmath inclusion tests. - Add driver test. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D61399/new/ https://reviews.llvm.org/D61399 Files: lib/Driver/Too

[PATCH] D61399: [OpenMP][Clang] Support for target math functions

gtbercea added a comment. @ABataev this patch works for both C and C++ and for both math.h and cmath headers. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D61399/new/ https://reviews.llvm.org/D61399 ___ cfe-commits

[PATCH] D61399: [OpenMP][Clang] Support for target math functions

gtbercea added a comment. In D61399#1489762 , @ABataev wrote: > In D61399#1489757 , @gtbercea wrote: > > > @ABataev this patch works for both C and C++ and for both math.h and cmath > > headers. > > > Did you test

[PATCH] D61399: [OpenMP][Clang] Support for target math functions

gtbercea updated this revision to Diff 198105. gtbercea added a comment. - Add driver test. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D61399/new/ https://reviews.llvm.org/D61399 Files: lib/Driver/ToolChains/Clang.cpp lib/Headers/CMakeLists.txt lib/Heade

[PATCH] D61399: [OpenMP][Clang] Support for target math functions

gtbercea updated this revision to Diff 198108. gtbercea added a comment. - Add new tests. Add stub headers. - Remove old tests. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D61399/new/ https://reviews.llvm.org/D61399 Files: lib/Driver/ToolChains/Clang.cpp li

[PATCH] D60568: [OpenMP] Add support for registering requires directives with the runtime

gtbercea marked an inline comment as done. gtbercea added inline comments. Comment at: lib/CodeGen/CGOpenMPRuntime.cpp:10364 + if (const auto *FD = dyn_cast(D)) { +if (OMPDeclareTargetDeclAttr::isDeclareTargetDeclaration(FD)) { + HasEmittedDeclareTargetRegion = true; --

[PATCH] D61399: [OpenMP][Clang] Support for target math functions

gtbercea marked 2 inline comments as done. gtbercea added inline comments. Comment at: lib/Headers/__clang_cuda_cmath.h:54 __DEVICE__ float fmod(float __x, float __y) { return ::fmodf(__x, __y); } +#ifndef _OPENMP __DEVICE__ int fpclassify(float __x) { ABataev

[PATCH] D61399: [OpenMP][Clang] Support for target math functions

gtbercea updated this revision to Diff 198301. gtbercea added a comment. - Address comments. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D61399/new/ https://reviews.llvm.org/D61399 Files: lib/Driver/ToolChains/Clang.cpp lib/Headers/CMakeLists.txt lib/Head

[PATCH] D61399: [OpenMP][Clang] Support for target math functions

gtbercea marked an inline comment as done. gtbercea added inline comments. Comment at: lib/Headers/__clang_cuda_cmath.h:444 using ::roundf; +#ifndef _OPENMP using ::scalblnf; ABataev wrote: > I see that the same guard is used > `lib/Headers/__clang_cuda_device

[PATCH] D61399: [OpenMP][Clang] Support for target math functions

gtbercea updated this revision to Diff 198311. gtbercea added a comment. - Address comments. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D61399/new/ https://reviews.llvm.org/D61399 Files: lib/Driver/ToolChains/Clang.cpp lib/Headers/CMakeLists.txt lib/Head

[PATCH] D61399: [OpenMP][Clang] Support for target math functions

gtbercea updated this revision to Diff 198578. gtbercea added a comment. - Fix move to openmp_wrapper folder. Fix header ordering problem. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D61399/new/ https://reviews.llvm.org/D61399 Files: lib/Driver/ToolChain.cpp

[PATCH] D61399: [OpenMP][Clang] Support for target math functions

gtbercea updated this revision to Diff 198664. gtbercea added a comment. - Eliminate declarations of functions not needed for math function resolution. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D61399/new/ https://reviews.llvm.org/D61399 Files: lib/Driver/T

[PATCH] D52434: [OpenMP] Make default schedules for NVPTX target regions in SPMD mode achieve coalescing

gtbercea created this revision. gtbercea added reviewers: ABataev, caomhin. Herald added subscribers: cfe-commits, guansong, jholewinski. For the OpenMP NVPTX toolchain choose default schedules which ensure coalescing on the GPU when in SPMD mode. This significantly increases the performance of

[PATCH] D52434: [OpenMP] Make default schedules for NVPTX target regions in SPMD mode achieve coalescing

gtbercea updated this revision to Diff 167172. gtbercea edited the summary of this revision. gtbercea added a comment. Only change default schedule for distribute directive. Repository: rC Clang https://reviews.llvm.org/D52434 Files: lib/CodeGen/CGOpenMPRuntime.cpp lib/CodeGen/CGOpenMPRu

[PATCH] D52434: [OpenMP] Make default schedules for NVPTX target regions in SPMD mode achieve coalescing

gtbercea updated this revision to Diff 167326. gtbercea added a comment. Fix type of chunk size. Repository: rC Clang https://reviews.llvm.org/D52434 Files: lib/CodeGen/CGOpenMPRuntime.h lib/CodeGen/CGOpenMPRuntimeNVPTX.cpp lib/CodeGen/CGOpenMPRuntimeNVPTX.h lib/CodeGen/CGStmtOpenMP.

[PATCH] D52434: [OpenMP] Make default schedules for NVPTX target regions in SPMD mode achieve coalescing

gtbercea added a comment. In https://reviews.llvm.org/D52434#1248012, @Hahnfeld wrote: > Should we also change the default `schedule` to `static, 1`? I know that's > not really needed for `teams distribute parallel for` (because the new > default `dist_schedule` only leaves one iteration per th

[PATCH] D52629: [OpenMP] Make default parallel for schedule in NVPTX target regions in SPMD mode achieve coalescing

gtbercea created this revision. gtbercea added reviewers: ABataev, Hahnfeld, caomhin. Herald added subscribers: cfe-commits, guansong, jholewinski. Set default schedule for parallel for loops to schedule(static, 1) when using SPMD mode on the NVPTX device offloading toolchain to ensure coalescing

[PATCH] D52629: [OpenMP] Make default parallel for schedule in NVPTX target regions in SPMD mode achieve coalescing

gtbercea updated this revision to Diff 167386. gtbercea added a comment. Address comment. Repository: rC Clang https://reviews.llvm.org/D52629 Files: lib/CodeGen/CGOpenMPRuntime.h lib/CodeGen/CGOpenMPRuntimeNVPTX.cpp lib/CodeGen/CGOpenMPRuntimeNVPTX.h lib/CodeGen/CGStmtOpenMP.cpp t

[PATCH] D52434: [OpenMP] Make default distribute schedule for NVPTX target regions in SPMD mode achieve coalescing

gtbercea added a comment. In https://reviews.llvm.org/D52434#1248844, @Hahnfeld wrote: > Just tested this and got very weird results for register usage: > > void func(double *a) { > #pragma omp target teams distribute parallel for map(a[0:100]) // > dist_schedule(static) > for (int i =

[PATCH] D52434: [OpenMP] Make default distribute schedule for NVPTX target regions in SPMD mode achieve coalescing

gtbercea added a comment. In https://reviews.llvm.org/D52434#1248975, @Hahnfeld wrote: > In https://reviews.llvm.org/D52434#1248974, @gtbercea wrote: > > > One big problem your code has is that the trip count is incredibly small, > > especially for STREAM and especially on GPUs. You need a much

[PATCH] D52434: [OpenMP] Make default distribute schedule for NVPTX target regions in SPMD mode achieve coalescing

gtbercea added a comment. In https://reviews.llvm.org/D52434#1248975, @Hahnfeld wrote: > In https://reviews.llvm.org/D52434#1248974, @gtbercea wrote: > > > One big problem your code has is that the trip count is incredibly small, > > especially for STREAM and especially on GPUs. You need a much

[PATCH] D52434: [OpenMP] Make default distribute schedule for NVPTX target regions in SPMD mode achieve coalescing

gtbercea added a comment. In https://reviews.llvm.org/D52434#1249186, @Hahnfeld wrote: > In https://reviews.llvm.org/D52434#1249102, @gtbercea wrote: > > > You report a slow down which I am not able to reproduce actually. Do you > > use any additional clauses not present in your previous post? >

[PATCH] D29660: [OpenMP] Add flag for overwriting default PTX version for OpenMP targets

gtbercea abandoned this revision. gtbercea added a comment. In https://reviews.llvm.org/D29660#1250333, @Hahnfeld wrote: > Going through my list of reviews, this patch was reverted because of memory > leaks in other changes. However, I don't think we need this anymore because > Clang is raising

[PATCH] D47394: [OpenMP][Clang][NVPTX] Replace bundling with partial linking for the OpenMP NVPTX device offloading toolchain

gtbercea updated this revision to Diff 187979. gtbercea added a comment. Herald added a subscriber: jdoerfert. Herald added a project: clang. - Update. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D47394/new/ https://reviews.llvm.org/D47394 Files: include/clan

[PATCH] D47394: [OpenMP][Clang][NVPTX] Replace bundling with partial linking for the OpenMP NVPTX device offloading toolchain

gtbercea added a comment. ping Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D47394/new/ https://reviews.llvm.org/D47394 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/l

[PATCH] D37912: [OpenMP] Bugfix: output file name drops the absolute path where full path is needed.

gtbercea created this revision. When composing the output file name, the path to the file is being dropped. The full path is required. Repository: rL LLVM https://reviews.llvm.org/D37912 Files: lib/Driver/ToolChains/Cuda.cpp Index: lib/Driver/ToolChains/Cuda.cpp

[PATCH] D37913: [OpenMP] Enable the existing nocudalib flag for OpenMP offloading toolchain.

gtbercea created this revision. Enable the -nocudalib flag for the OpenMP device offloading toolchain as well. Currently it can only be used for the CUDA toolchain. Repository: rL LLVM https://reviews.llvm.org/D37913 Files: lib/Driver/ToolChains/Cuda.cpp Index: lib/Driver/ToolChains/Cud

[PATCH] D37914: [OpenMP] Don't throw cudalib not found error if only front-end is required.

gtbercea created this revision. If we only use the compiler front-end, do not throw an error about the cuda device library not being found. This allows the front-end to be run on systems where no Cuda installation is found. Repository: rL LLVM https://reviews.llvm.org/D37914 Files: lib/D

[PATCH] D37912: [OpenMP] Bugfix: output file name drops the absolute path where full path is needed.

gtbercea added a comment. In https://reviews.llvm.org/D37912#872294, @tra wrote: > Shouldn't this temp .cubin file go into the temporary directory, as opposed > to the same directory as the input file? That is indeed the intention. The filename already contains the "/tmp/" I just make sure th

[PATCH] D37914: [OpenMP] Don't throw cudalib not found error if only front-end is required.

gtbercea updated this revision to Diff 115444. gtbercea added a comment. Contract check. https://reviews.llvm.org/D37914 Files: lib/Driver/ToolChains/Cuda.cpp Index: lib/Driver/ToolChains/Cuda.cpp === --- lib/Driver/ToolChains/

[PATCH] D37914: [OpenMP] Don't throw cudalib not found error if only front-end is required.

gtbercea updated this revision to Diff 115447. gtbercea added a comment. Fix parantheses. https://reviews.llvm.org/D37914 Files: lib/Driver/ToolChains/Cuda.cpp Index: lib/Driver/ToolChains/Cuda.cpp === --- lib/Driver/ToolChains

[PATCH] D37914: [OpenMP] Don't throw cudalib not found error if only front-end is required.

gtbercea updated this revision to Diff 115481. gtbercea added a comment. Fix condition. https://reviews.llvm.org/D37914 Files: lib/Driver/ToolChains/Cuda.cpp Index: lib/Driver/ToolChains/Cuda.cpp === --- lib/Driver/ToolChains/C

[PATCH] D37912: [OpenMP] Bugfix: output file name drops the absolute path where full path is needed.

gtbercea updated this revision to Diff 115497. gtbercea added a comment. Fix tests. https://reviews.llvm.org/D37912 Files: lib/Driver/ToolChains/Cuda.cpp test/Driver/openmp-offload-gpu.c Index: test/Driver/openmp-offload-gpu.c ==

[PATCH] D37914: [OpenMP] Don't throw cudalib not found error if only front-end is required.

gtbercea updated this revision to Diff 115500. gtbercea added a comment. Add test. https://reviews.llvm.org/D37914 Files: lib/Driver/ToolChains/Cuda.cpp test/Driver/openmp-offload-gpu.c Index: test/Driver/openmp-offload-gpu.c ===

[PATCH] D37914: [OpenMP] Don't throw cudalib not found error if only front-end is required.

gtbercea updated this revision to Diff 115502. gtbercea added a comment. Fix diff. Repository: rL LLVM https://reviews.llvm.org/D37914 Files: lib/Driver/ToolChains/Cuda.cpp test/Driver/openmp-offload-gpu.c Index: test/Driver/openmp-offload-gpu.c

[PATCH] D37912: [OpenMP] Bugfix: output file name drops the absolute path where full path is needed.

gtbercea updated this revision to Diff 115667. gtbercea added a reviewer: hfinkel. Repository: rL LLVM https://reviews.llvm.org/D37912 Files: lib/Driver/ToolChains/Cuda.cpp test/Driver/openmp-offload-gpu.c Index: test/Driver/openmp-offload-gpu.c ==

[PATCH] D37913: [OpenMP] Enable the existing nocudalib flag for OpenMP offloading toolchain.

gtbercea updated this revision to Diff 115692. gtbercea added a comment. Add test. Repository: rL LLVM https://reviews.llvm.org/D37913 Files: lib/Driver/ToolChains/Cuda.cpp test/Driver/openmp-offload-gpu.c Index: test/Driver/openmp-offload-gpu.c

[PATCH] D37914: [OpenMP] Don't throw cudalib not found error if only front-end is required.

gtbercea updated this revision to Diff 115697. gtbercea added a comment. Only check for -S. Repository: rL LLVM https://reviews.llvm.org/D37914 Files: lib/Driver/ToolChains/Cuda.cpp test/Driver/openmp-offload-gpu.c Index: test/Driver/openmp-offload-gpu.c ===

[PATCH] D37914: [OpenMP] Don't throw cudalib not found error if only front-end is required.

gtbercea added inline comments. Comment at: test/Driver/openmp-offload-gpu.c:133 +/// Check that the flag is passed when -fopenmp-relocatable-target is used. +// RUN: %clang -### -c -S -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda -Xopenmp-target -march=sm_99 -fopenmp-

[PATCH] D38040: [OpenMP] Add an additional test for D34888

gtbercea created this revision. Test for checking if the mapping is performed correctly. This is a test initially included in Patch https://reviews.llvm.org/D29905 Repository: rL LLVM https://reviews.llvm.org/D38040 Files: test/OpenMP/target_map_codegen.cpp Index: test/OpenMP/target_map

[PATCH] D37912: [OpenMP] Bugfix: output file name drops the absolute path where full path is needed.

gtbercea updated this revision to Diff 115941. gtbercea added a comment. Address comment. Repository: rL LLVM https://reviews.llvm.org/D37912 Files: lib/Driver/ToolChains/Cuda.cpp test/Driver/openmp-offload-gpu.c Index: test/Driver/openmp-offload-gpu.c =

[PATCH] D37913: [OpenMP] Enable the existing nocudalib flag for OpenMP offloading toolchain.

gtbercea updated this revision to Diff 115943. gtbercea added a comment. Don't take into account unknown CUDA archs not even for testing purposes. Repository: rL LLVM https://reviews.llvm.org/D37913 Files: lib/Driver/ToolChains/Cuda.cpp test/Driver/openmp-offload-gpu.c Index: test/Driv

[PATCH] D38040: [OpenMP] Add an additional test for D34888

gtbercea updated this revision to Diff 115947. Repository: rL LLVM https://reviews.llvm.org/D38040 Files: test/OpenMP/target_map_codegen.cpp Index: test/OpenMP/target_map_codegen.cpp === --- test/OpenMP/target_map_codegen.cpp

[PATCH] D37914: [OpenMP] Don't throw cudalib not found error if only front-end is required.

gtbercea updated this revision to Diff 115948. Repository: rL LLVM https://reviews.llvm.org/D37914 Files: lib/Driver/ToolChains/Cuda.cpp test/Driver/openmp-offload-gpu.c Index: test/Driver/openmp-offload-gpu.c === --- test/D

[PATCH] D38040: [OpenMP] Add an additional test for D34888

gtbercea added a comment. In https://reviews.llvm.org/D38040#878090, @Hahnfeld wrote: > Hi Doru, > > if I remember correctly I submitted https://reviews.llvm.org/D34888 for a > crash when mapping a scalar value with nested regions. > I've marked another test in this file that the codegen for `t

[PATCH] D38040: [OpenMP] Add an additional test for D34888

gtbercea added a comment. In https://reviews.llvm.org/D38040#878799, @Hahnfeld wrote: > In https://reviews.llvm.org/D38040#878441, @gtbercea wrote: > > > The test is verifying whether the parameter is passed to the kernel > > correctly. I believe it was not passed as a reference before the patch

[PATCH] D59424: [OpenMP][NVPTX] Replace void** buffer by byte-wise buffer

gtbercea added a comment. Could you check if there is any change in the number of registers new scheme vs. old scheme? Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D59424/new/ https://reviews.llvm.org/D59424

[PATCH] D59319: [OpenMP][Offloading][1/3] A generic and simple target region interface

gtbercea added a comment. Could you add some tests for this? Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D59319/new/ https://reviews.llvm.org/D59319 ___ cfe-commits mailing list cfe-commits@lists.llvm

[PATCH] D59319: [OpenMP][Offloading][1/3] A generic and simple target region interface

gtbercea added a comment. Could you check what the difference is between the same kernel in today's SPMD mode vs the SPMD mode produced via this method? Number of registers, instructions, checking everything gets optimized out as expected. The LLVM-IR should be almost identical. Repository:

[PATCH] D63009: [OpenMP] Add target task alloc function with device ID

gtbercea created this revision. gtbercea added reviewers: ABataev, AlexEichenberger, caomhin. Herald added subscribers: cfe-commits, jdoerfert, guansong. Herald added a project: clang. Add a new call to Clang to perform task allocation for the target. Repository: rC Clang https://reviews.llvm

[PATCH] D63108: [OpenMP] Add support for handling declare target to clause when unified memory is required

gtbercea created this revision. gtbercea added reviewers: ABataev, AlexEichenberger, caomhin. Herald added subscribers: cfe-commits, jdoerfert, guansong. Herald added a project: clang. This patch adds support for the handling of the variables under the declare target to clause. The variables in

[PATCH] D63010: [OpenMP] Add task alloc function

gtbercea updated this revision to Diff 204138. gtbercea added a comment. Herald added a project: clang. Herald added a subscriber: cfe-commits. - Add tests. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D63010/new/ https://reviews.llvm.org/D63010 Files: lib/Cod

[PATCH] D63009: [OpenMP] Add target task alloc function with device ID

gtbercea updated this revision to Diff 204139. gtbercea added a comment. Herald added a project: OpenMP. Herald added a subscriber: openmp-commits. - Add temporary implementation. Repository: rOMP OpenMP CHANGES SINCE LAST ACTION https://reviews.llvm.org/D63009/new/ https://reviews.llvm.or

[PATCH] D63010: [OpenMP] Add task alloc function

gtbercea updated this revision to Diff 204140. gtbercea added a comment. - Add temporary implementation. Repository: rOMP OpenMP CHANGES SINCE LAST ACTION https://reviews.llvm.org/D63010/new/ https://reviews.llvm.org/D63010 Files: runtime/src/kmp.h runtime/src/kmp_tasking.cpp Index:

[PATCH] D63009: [OpenMP] Add target task alloc function with device ID

gtbercea updated this revision to Diff 204141. gtbercea added a comment. - Add tests. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D63009/new/ https://reviews.llvm.org/D63009 Files: lib/CodeGen/CGOpenMPRuntime.cpp test/OpenMP/target_depend_codegen.cpp test

[PATCH] D63009: [OpenMP] Add target task alloc function with device ID

gtbercea updated this revision to Diff 204294. gtbercea added a comment. - Add device ID if available. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D63009/new/ https://reviews.llvm.org/D63009 Files: lib/CodeGen/CGOpenMPRuntime.cpp test/OpenMP/target_depend_c

[PATCH] D63009: [OpenMP] Add target task alloc function with device ID

gtbercea updated this revision to Diff 204320. gtbercea added a comment. - Fix tests. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D63009/new/ https://reviews.llvm.org/D63009 Files: lib/CodeGen/CGOpenMPRuntime.cpp test/OpenMP/target_depend_codegen.cpp test

[PATCH] D63009: [OpenMP] Add target task alloc function with device ID

gtbercea updated this revision to Diff 204763. gtbercea added a comment. - Fix function name. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D63009/new/ https://reviews.llvm.org/D63009 Files: lib/CodeGen/CGOpenMPRuntime.cpp test/OpenMP/target_depend_codegen.cp

[PATCH] D63009: [OpenMP] Add target task alloc function with device ID

gtbercea marked 2 inline comments as done. gtbercea added inline comments. Comment at: lib/CodeGen/CGOpenMPRuntime.cpp:5122 +NewTask = CGF.EmitRuntimeCall( + createRuntimeFunction(OMPRTL__kmpc_omp_target_task_alloc), AllocArgs); + } else { ABataev wrote

[PATCH] D63009: [OpenMP] Add target task alloc function with device ID

gtbercea marked an inline comment as done. gtbercea added inline comments. Comment at: lib/CodeGen/CGOpenMPRuntime.cpp:5122 +NewTask = CGF.EmitRuntimeCall( + createRuntimeFunction(OMPRTL__kmpc_omp_target_task_alloc), AllocArgs); + } else { ABataev wrote

[PATCH] D63009: [OpenMP] Add target task alloc function with device ID

gtbercea marked an inline comment as done. gtbercea added inline comments. Comment at: lib/CodeGen/CGOpenMPRuntime.cpp:5122 +NewTask = CGF.EmitRuntimeCall( + createRuntimeFunction(OMPRTL__kmpc_omp_target_task_alloc), AllocArgs); + } else { ABataev wrote

[PATCH] D63009: [OpenMP] Add target task alloc function with device ID

gtbercea marked an inline comment as done. gtbercea added inline comments. Comment at: lib/CodeGen/CGOpenMPRuntime.cpp:5122 +NewTask = CGF.EmitRuntimeCall( + createRuntimeFunction(OMPRTL__kmpc_omp_target_task_alloc), AllocArgs); + } else { ABataev wrote

[PATCH] D63108: [OpenMP] Add support for handling declare target to clause when unified memory is required

gtbercea added a comment. ping Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D63108/new/ https://reviews.llvm.org/D63108 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/l

[PATCH] D60883: [OpenMP] Avoid emitting maps for target link variables when unified memory is used

gtbercea updated this revision to Diff 204780. gtbercea added a comment. - Remove virtual. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D60883/new/ https://reviews.llvm.org/D60883 Files: lib/CodeGen/CGOpenMPRuntime.cpp lib/CodeGen/CGOpenMPRuntime.h lib/Sem

[PATCH] D60883: [OpenMP] Avoid emitting maps for target link variables when unified memory is used

gtbercea updated this revision to Diff 204806. gtbercea added a comment. - Make function const. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D60883/new/ https://reviews.llvm.org/D60883 Files: lib/CodeGen/CGOpenMPRuntime.cpp lib/CodeGen/CGOpenMPRuntime.h li

[PATCH] D63010: [OpenMP] Add task alloc function

gtbercea added a comment. ping Repository: rOMP OpenMP CHANGES SINCE LAST ACTION https://reviews.llvm.org/D63010/new/ https://reviews.llvm.org/D63010 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailma

[PATCH] D63009: [OpenMP] Add target task alloc function with device ID

gtbercea marked an inline comment as done. gtbercea added inline comments. Comment at: lib/CodeGen/CGOpenMPRuntime.cpp:5122 +NewTask = CGF.EmitRuntimeCall( + createRuntimeFunction(OMPRTL__kmpc_omp_target_task_alloc), AllocArgs); + } else { AlexEichenber

[PATCH] D63009: [OpenMP] Add target task alloc function with device ID

gtbercea added a comment. In D63009#1544784 , @ABataev wrote: > The tests must check the device ID for target-based calls of the task alloc > function. Since device ID is now an expression I can only do this: i64 {{.*}}) Is this what you wanted? @AB

[PATCH] D63009: [OpenMP] Add target task alloc function with device ID

gtbercea added a comment. In D63009#1544758 , @Hahnfeld wrote: > Am I correct that the second to last revision ("- Fix tests.") removed all > checks for the actual `device_id` argument from the tests? From my point of > view that's not fixing but weakeni

[PATCH] D63009: [OpenMP] Add target task alloc function with device ID

gtbercea added a comment. In D63009#1544984 , @ABataev wrote: > In D63009#1544900 , @gtbercea wrote: > > > In D63009#1544758 , @Hahnfeld > > wrote: > > > > > Am I correct th

[PATCH] D63454: [OpenMP] Strengthen regression tests for task allocation under nowait depend clauses NFC

gtbercea created this revision. gtbercea added reviewers: ABataev, Hahnfeld, caomhin. Herald added subscribers: cfe-commits, jdoerfert, guansong. Herald added a project: clang. This patch strengthens the tests introduced in D63009 by: - adding new test for defau

[PATCH] D63454: [OpenMP] Strengthen regression tests for task allocation under nowait depend clauses NFC

gtbercea updated this revision to Diff 205186. gtbercea added a comment. - Fix test. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D63454/new/ https://reviews.llvm.org/D63454 Files: test/OpenMP/target_constant_device_codegen.cpp test/OpenMP/target_depend_code

[PATCH] D63454: [OpenMP] Strengthen regression tests for task allocation under nowait depend clauses NFC

gtbercea added a comment. ping Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D63454/new/ https://reviews.llvm.org/D63454 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/l

[PATCH] D63009: [OpenMP] Add target task alloc function with device ID

gtbercea added a comment. Fixed in D63454 Repository: rL LLVM CHANGES SINCE LAST ACTION https://reviews.llvm.org/D63009/new/ https://reviews.llvm.org/D63009 ___ cfe-commits mailing list cfe-commits@lists.llvm.

[PATCH] D63108: [OpenMP] Add support for handling declare target to clause when unified memory is required

gtbercea updated this revision to Diff 205667. gtbercea marked an inline comment as done. gtbercea added a comment. - Merge MT_Link and MT_To with unified memory cases. - Transform switch into if statements. - Fix declare target attribute checks. Repository: rC Clang CHANGES SINCE LAST ACTION

[PATCH] D63108: [OpenMP] Add support for handling declare target to clause when unified memory is required

gtbercea added a comment. Still need to update the test but the rest of the code is updated. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D63108/new/ https://reviews.llvm.org/D63108 ___ cfe-commits mailing list cfe-c

[PATCH] D63108: [OpenMP] Add support for handling declare target to clause when unified memory is required

gtbercea updated this revision to Diff 205825. gtbercea added a comment. - Address comments. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D63108/new/ https://reviews.llvm.org/D63108 Files: lib/CodeGen/CGDeclCXX.cpp lib/CodeGen/CGExpr.cpp lib/CodeGen/CGOpen

[PATCH] D63108: [OpenMP] Add support for handling declare target to clause when unified memory is required

gtbercea updated this revision to Diff 205839. gtbercea added a comment. - Address comments. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D63108/new/ https://reviews.llvm.org/D63108 Files: lib/CodeGen/CGDeclCXX.cpp lib/CodeGen/CGExpr.cpp lib/CodeGen/CGOpen

[PATCH] D64592: [OpenMP] Fix unified memory implementation for multiple compilation units

gtbercea created this revision. gtbercea added reviewers: ABataev, jdoerfert, caomhin. Herald added subscribers: cfe-commits, guansong. Herald added a project: clang. This patch fixes the case where variables in different compilation units have the same name. Repository: rC Clang https://rev

[PATCH] D64592: [OpenMP] Fix unified memory implementation for multiple compilation units

gtbercea updated this revision to Diff 210386. gtbercea added a comment. - Fix tests. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D64592/new/ https://reviews.llvm.org/D64592 Files: lib/CodeGen/CGOpenMPRuntime.cpp test/OpenMP/declare_target_codegen.cpp tes

[PATCH] D64592: [OpenMP] Fix unified memory implementation for multiple compilation units

gtbercea marked an inline comment as done. gtbercea added inline comments. Comment at: lib/CodeGen/CGOpenMPRuntime.cpp:2610 + auto *GV = cast(Ptr); + GV->setLinkage(llvm::GlobalValue::WeakAnyLinkage); + ABataev wrote: > Better to fix the link clause pro

[PATCH] D64878: [OpenMP] Fix sema check for unified memory case NFC

gtbercea created this revision. gtbercea added a reviewer: ABataev. Herald added subscribers: cfe-commits, jdoerfert, guansong. Herald added a project: clang. This patch fixes a condition introduced in patch D60883 . Repository: rC Clang https://reviews.llvm.

[PATCH] D64926: [OpenMP] Fx link clause

gtbercea created this revision. gtbercea added reviewers: ABataev, caomhin. Herald added subscribers: cfe-commits, jdoerfert, guansong. Herald added a project: clang. Adjust linkage of variables under the link clause. Repository: rC Clang https://reviews.llvm.org/D64926 Files: lib/CodeGen/

<    1   2   3   4   5   6   7   >