Skip to content

GitLab

Commit e704bbc8 authored by wenjh's avatar wenjh
Browse files

Fix build error 


Signed-off-by: wenjh's avatarwenjh <wenjh@sugon.com>
parent 331f2fc4
Hide whitespace changes
Inline Side-by-side
Showing with 2 additions and 2 deletions
transformer_engine/common/transpose/quantize_transpose_vector_blockwise.cu
View file @ e704bbc8
...@@ -505,7 +505,7 @@ __global__ void __launch_bounds__(kThreadsPerBlock) block_scaled_1d_cast_transpo ...@@ -505,7 +505,7 @@ __global__ void __launch_bounds__(kThreadsPerBlock) block_scaled_1d_cast_transpo
amax = warp_reduce_max<kThreadsPerWarp>(amax); amax = warp_reduce_max<kThreadsPerWarp>(amax);
constexpr int lane_zero = 0; constexpr int lane_zero = 0;
#ifdef __HIP_PLATFORM_AMD__ #ifdef __HIP_PLATFORM_AMD__
amax = __shfl_sync(0xFFFFFFFF, amax, lane_zero, kThreadsPerWarp); amax = __shfl_sync((unsigned long long)(0xFFFFFFFF), amax, lane_zero, kThreadsPerWarp);
#else #else
amax = __shfl_sync(0xFFFFFFFF, amax, lane_zero); amax = __shfl_sync(0xFFFFFFFF, amax, lane_zero);
#endif #endif
...@@ -895,7 +895,7 @@ __global__ void __launch_bounds__(kThreadsPerBlock) block_scaled_1d_cast_transpo ...@@ -895,7 +895,7 @@ __global__ void __launch_bounds__(kThreadsPerBlock) block_scaled_1d_cast_transpo
amax = warp_reduce_max<kThreadsPerWarp>(amax); amax = warp_reduce_max<kThreadsPerWarp>(amax);
constexpr int lane_zero = 0; constexpr int lane_zero = 0;
#ifdef __HIP_PLATFORM_AMD__ #ifdef __HIP_PLATFORM_AMD__
amax = __shfl_sync(0xFFFFFFFF, amax, lane_zero, kThreadsPerWarp); amax = __shfl_sync((unsigned long long)(0xFFFFFFFF), amax, lane_zero, kThreadsPerWarp);
#else #else
amax = __shfl_sync(0xFFFFFFFF, amax, lane_zero); amax = __shfl_sync(0xFFFFFFFF, amax, lane_zero);
#endif #endif
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment