Commit c3946bf0 authored by wooway777's avatar wooway777
Browse files

issue/1032 - fix block size on iluvatar

parent abd45713
...@@ -106,10 +106,12 @@ infiniStatus_t Descriptor::calculate( ...@@ -106,10 +106,12 @@ infiniStatus_t Descriptor::calculate(
return INFINI_STATUS_BAD_TENSOR_DTYPE; \ return INFINI_STATUS_BAD_TENSOR_DTYPE; \
} }
if (_opaque->internal->maxThreadsPerBlock() == CUDA_BLOCK_SIZE_1024) { if (_opaque->internal->maxThreadsPerBlock() == CUDA_BLOCK_SIZE_512) {
CALCULATE_SWIGLU_CUDA_WITH_BLOCK_SIZE(CUDA_BLOCK_SIZE_1024)
} else if (_opaque->internal->maxThreadsPerBlock() == CUDA_BLOCK_SIZE_512) {
CALCULATE_SWIGLU_CUDA_WITH_BLOCK_SIZE(CUDA_BLOCK_SIZE_512) CALCULATE_SWIGLU_CUDA_WITH_BLOCK_SIZE(CUDA_BLOCK_SIZE_512)
} else if (_opaque->internal->maxThreadsPerBlock() == CUDA_BLOCK_SIZE_1024) {
CALCULATE_SWIGLU_CUDA_WITH_BLOCK_SIZE(CUDA_BLOCK_SIZE_1024)
} else if (_opaque->internal->maxThreadsPerBlock() == CUDA_BLOCK_SIZE_2048) {
CALCULATE_SWIGLU_CUDA_WITH_BLOCK_SIZE(CUDA_BLOCK_SIZE_2048)
} else if (_opaque->internal->maxThreadsPerBlock() == CUDA_BLOCK_SIZE_4096) { } else if (_opaque->internal->maxThreadsPerBlock() == CUDA_BLOCK_SIZE_4096) {
CALCULATE_SWIGLU_CUDA_WITH_BLOCK_SIZE(CUDA_BLOCK_SIZE_4096) CALCULATE_SWIGLU_CUDA_WITH_BLOCK_SIZE(CUDA_BLOCK_SIZE_4096)
} else { } else {
......
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