[PATCH] D95313: [WIP] Move part of nvptx devicertl under clang

[Jon Chesterfield via Phabricator via cfe-commits]

JonChesterfield abandoned this revision.
JonChesterfield added a comment.

Abandoned in favour of multiple instantiations of the devicertl, which works 
across all languages and doesn't require hacks to clang


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D95313/new/

https://reviews.llvm.org/D95313

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D95313: [WIP] Move part of nvptx devicertl under clang

[Shilei Tian via Phabricator via cfe-commits]

tianshilei1992 added a comment.

In general we're moving to the direction that target specific implementation 
will be compiled along with user code, which is fantastic. In this way, we only 
need to provide one bitcode library for one target. The change in FE lacks of 
some efficiency. If user code has multiple files, target specific header will 
be included multiple times, thus compiled multiple times. A more efficient way 
is to change the workflow of the driver, probably in the following way:

1. Compile target implementation `t.bc`
2. Link `t.bc` and `libomptarget-[arch].bc` to `libomptarget.bc`
3. Compile user code, which is also multiple steps. `libomptarget.bc` is fed 
into FE in this step.
4. Remaining steps...




Comment at: clang/lib/Driver/ToolChains/Clang.cpp:1204
+{
+  auto *CTC = static_cast(
+  C.getSingleOffloadToolChain());

JonChesterfield wrote:
> Logic very like this could pick out a second, small devicertl bitcode library
can we just use one header with different macros, like what we're using now?


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D95313/new/

https://reviews.llvm.org/D95313

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D95313: [WIP] Move part of nvptx devicertl under clang

[Jon Chesterfield via Phabricator via cfe-commits]

JonChesterfield added inline comments.



Comment at: clang/lib/Driver/ToolChains/Clang.cpp:1204
+{
+  auto *CTC = static_cast(
+  C.getSingleOffloadToolChain());

Logic very like this could pick out a second, small devicertl bitcode library



Comment at: 
clang/lib/Headers/openmp_wrappers/__clang_openmp_devicertl_cuda_ge90.h:18
+#ifdef _OPENMP
+#define DEVICE __attribute__((used))
+#else

linkonce and linkonce_odr can both be discarded, but these symbols need to 
survive until devicertl is linked



Comment at: 
clang/lib/Headers/openmp_wrappers/__clang_openmp_devicertl_cuda_ge90.h:39
+  int WARPSIZE = 32;
+  int tmp = ((WARPSIZE - Width) << 8) | 0x1f;
+  return __nvvm_shfl_sync_down_i32(Mask, Var, Delta, tmp);

calling into intrinsics would remove this messy expression, if we can work out 
how to reliably compile parts of the cuda sdk as openmp

these functions probably can't be instantiated on the host, so when changing 
the devicertl build over to openmp we will also need to guard these with 
variant / macro



Comment at: openmp/libomptarget/deviceRTLs/amdgcn/src/target_impl.hip:49
 // Warp vote function
-DEVICE __kmpc_impl_lanemask_t __kmpc_impl_activemask() {
+EXTERN __kmpc_impl_lanemask_t __kmpc_impl_activemask() {
   return __builtin_amdgcn_read_exec();

Changing from c++ to c name mangling seems fine here. We don't consistently use 
one or the other in the devicertl. Using c mangling is simpler if the 
implementation is in a header.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D95313/new/

https://reviews.llvm.org/D95313

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D95313: [WIP] Move part of nvptx devicertl under clang

[Jon Chesterfield via Phabricator via cfe-commits]

JonChesterfield created this revision.
JonChesterfield added reviewers: jdoerfert, tianshilei1992.
Herald added subscribers: mgorny, jvesely.
JonChesterfield requested review of this revision.
Herald added subscribers: openmp-commits, cfe-commits, sstefan1.
Herald added projects: clang, OpenMP.

[WIP] Move part of nvptx devicertl under clang

Example of moving the devicertl functions that depend on cuda
version under clang, so they can be injected at application
build time.

The original idea was to use the intrinsic definitions from
__clang_cuda_intrinsics, but that header needs a lot of cuda
specific setup to compile and includes part of the cuda sdk.
It's therefore difficult to compile as openmp.

This implements the code in headers and will work for c++ with
openmp, but not necessarily for C as the inline functions may not
be instantiated. It will also be a problem for fortran openmp.

I'm inclined to do something broadly equivalent to this, but in
the library. It means clang would need to link against devicertl.bc
and against a small cuda version specific devicertl_tbd.bc.


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D95313

Files:
  clang/lib/Driver/ToolChains/Clang.cpp
  clang/lib/Headers/CMakeLists.txt
  clang/lib/Headers/openmp_wrappers/__clang_openmp_devicertl_cuda_ge90.h
  clang/lib/Headers/openmp_wrappers/__clang_openmp_devicertl_cuda_lt90.h
  openmp/libomptarget/deviceRTLs/amdgcn/src/target_impl.h
  openmp/libomptarget/deviceRTLs/amdgcn/src/target_impl.hip
  openmp/libomptarget/deviceRTLs/nvptx/src/target_impl.cu
  openmp/libomptarget/deviceRTLs/nvptx/src/target_impl.h

Index: openmp/libomptarget/deviceRTLs/nvptx/src/target_impl.h
===
--- openmp/libomptarget/deviceRTLs/nvptx/src/target_impl.h
+++ openmp/libomptarget/deviceRTLs/nvptx/src/target_impl.h
@@ -100,17 +100,18 @@
 #error CUDA_VERSION macro is undefined, something wrong with cuda.
 #endif
 
-DEVICE __kmpc_impl_lanemask_t __kmpc_impl_activemask();
+EXTERN __kmpc_impl_lanemask_t __kmpc_impl_activemask();
 
-DEVICE int32_t __kmpc_impl_shfl_sync(__kmpc_impl_lanemask_t Mask, int32_t Var,
+EXTERN int32_t __kmpc_impl_shfl_sync(__kmpc_impl_lanemask_t Mask, int32_t Var,
  int32_t SrcLane);
 
-DEVICE int32_t __kmpc_impl_shfl_down_sync(__kmpc_impl_lanemask_t Mask,
+EXTERN int32_t __kmpc_impl_shfl_down_sync(__kmpc_impl_lanemask_t Mask,
   int32_t Var, uint32_t Delta,
   int32_t Width);
 
+EXTERN void __kmpc_impl_syncwarp(__kmpc_impl_lanemask_t Mask);
+
 DEVICE void __kmpc_impl_syncthreads();
-DEVICE void __kmpc_impl_syncwarp(__kmpc_impl_lanemask_t Mask);
 
 // NVPTX specific kernel initialization
 DEVICE void __kmpc_impl_target_init();
Index: openmp/libomptarget/deviceRTLs/nvptx/src/target_impl.cu
===
--- openmp/libomptarget/deviceRTLs/nvptx/src/target_impl.cu
+++ openmp/libomptarget/deviceRTLs/nvptx/src/target_impl.cu
@@ -18,17 +18,6 @@
 
 // Forward declaration of CUDA primitives which will be evetually transformed
 // into LLVM intrinsics.
-extern "C" {
-unsigned int __activemask();
-unsigned int __ballot(unsigned);
-// The default argument here is based on NVIDIA's website
-// https://developer.nvidia.com/blog/using-cuda-warp-level-primitives/
-int __shfl_sync(unsigned mask, int val, int src_line, int width = WARPSIZE);
-int __shfl(int val, int src_line, int width = WARPSIZE);
-int __shfl_down(int var, unsigned detla, int width);
-int __shfl_down_sync(unsigned mask, int var, unsigned detla, int width);
-void __syncwarp(int mask);
-}
 
 DEVICE void __kmpc_impl_unpack(uint64_t val, uint32_t , uint32_t ) {
   asm volatile("mov.b64 {%0,%1}, %2;" : "=r"(lo), "=r"(hi) : "l"(val));
@@ -69,45 +58,8 @@
   return (double)nsecs * __kmpc_impl_get_wtick();
 }
 
-// In Cuda 9.0, __ballot(1) from Cuda 8.0 is replaced with __activemask().
-DEVICE __kmpc_impl_lanemask_t __kmpc_impl_activemask() {
-#if CUDA_VERSION >= 9000
-  return __activemask();
-#else
-  return __ballot(1);
-#endif
-}
-
-// In Cuda 9.0, the *_sync() version takes an extra argument 'mask'.
-DEVICE int32_t __kmpc_impl_shfl_sync(__kmpc_impl_lanemask_t Mask, int32_t Var,
- int32_t SrcLane) {
-#if CUDA_VERSION >= 9000
-  return __shfl_sync(Mask, Var, SrcLane);
-#else
-  return __shfl(Var, SrcLane);
-#endif // CUDA_VERSION
-}
-
-DEVICE int32_t __kmpc_impl_shfl_down_sync(__kmpc_impl_lanemask_t Mask,
-  int32_t Var, uint32_t Delta,
-  int32_t Width) {
-#if CUDA_VERSION >= 9000
-  return __shfl_down_sync(Mask, Var, Delta, Width);
-#else
-  return __shfl_down(Var, Delta, Width);
-#endif // CUDA_VERSION
-}
-
 DEVICE void __kmpc_impl_syncthreads() { __syncthreads(); }
 
-DEVICE void __kmpc_impl_syncwarp(__kmpc_impl_lanemask_t