Skip to content

Some fixes #8

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
justinkb wants to merge 2 commits into
base: rocm
Choose a base branch
from
Open

Some fixes #8

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
3477131
fix warp size define
justinkb Jan 25, 2025
afcc4b3
fix argument order
justinkb Jan 25, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 2 additions & 1 deletion src/cuda/helpers.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,6 +6,7 @@
#ifdef CT2_USE_HIP
#include <hip/hip_fp16.h>
#include <hip/hip_bf16.h>
#include <hip/hip_runtime.h>

#define __nv_bfloat16 __hip_bfloat16
__device__ inline void __syncwarp(uint32_t mask){} //TODO: 6.1 should have this but it doesn't?
Expand Down Expand Up @@ -414,7 +415,7 @@ namespace ctranslate2 {
// They help define row-wise reduction where each block handles a single row.

#ifdef CT2_USE_HIP
#define C10_WARP_SIZE 64 //TODO: detect arch to set 32 for rdna
#define C10_WARP_SIZE warpSize
#else
#define C10_WARP_SIZE 32
#endif
Expand Down
2 changes: 1 addition & 1 deletion src/ops/conv1d_gpu.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -67,8 +67,8 @@ namespace ctranslate2 {
size_t workspace_size = 0;
void* workspace = nullptr;
CUDNN_CHECK(miopenConvolutionForwardGetWorkSpaceSize(handle,
input_desc,
weight_desc,
input_desc,
conv_desc,
output_desc,
&workspace_size));
Expand Down