[Openmp-commits] [PATCH] D138389: [OpenMP][libomptarget] Add AMDGPU NextGen plugin with asynchronous behavior

[Ye Luo via Phabricator via Openmp-commits]

ye-luo added inline comments.


================
Comment at: openmp/libomptarget/plugins-nextgen/amdgpu/src/rtl.cpp:1461
+        OMPX_NumQueues("LIBOMPTARGET_AMDGPU_NUM_QUEUES", 8),
+        OMPX_QueueSize("LIBOMPTARGET_AMDGPU_QUEUE_SIZE", 1024),
+        OMPX_StreamSize("LIBOMPTARGET_AMDGPU_STREAM_SIZE", 512),
----------------
jdoerfert wrote:
> 4 x 512 x 128 might be enough
What is QUEUE_SIZE? Prefer not to use SIZE but NUM_XXX_PER_QUEUE


================
Comment at: openmp/libomptarget/plugins-nextgen/amdgpu/src/rtl.cpp:1462
+        OMPX_QueueSize("LIBOMPTARGET_AMDGPU_QUEUE_SIZE", 1024),
+        OMPX_StreamSize("LIBOMPTARGET_AMDGPU_STREAM_SIZE", 512),
+        OMPX_MaxAsyncCopySize("LIBOMPTARGET_AMDGPU_MAX_ASYNC_COPY_SIZE",
----------------
What is the STREAM_SIZE?


================
Comment at: openmp/libomptarget/plugins-nextgen/amdgpu/src/rtl.cpp:1463
+        OMPX_StreamSize("LIBOMPTARGET_AMDGPU_STREAM_SIZE", 512),
+        OMPX_MaxAsyncCopySize("LIBOMPTARGET_AMDGPU_MAX_ASYNC_COPY_SIZE",
+                              1 * 1024 * 1024), // 1MB
----------------
LIBOMPTARGET_AMDGPU_MAX_ASYNC_COPY_BYTES?


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D138389/new/

https://reviews.llvm.org/D138389