dnrops
/
cutlass
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Fixed illegal PTX syntax (#1225)

This commit is contained in:
Haicheng Wu 2023-12-01 12:29:48 -05:00 committed by GitHub
parent bef1fbcbe6
commit 4a1709e17e
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 2 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
include/cutlass/arch/barrier.h
View File

@ -267,13 +267,12 @@ public:
static void arrive(ValueType const* smem_ptr) {
#if CUDA_BARRIER_ENABLED
uint32_t smem_addr = cute::cast_smem_ptr_to_uint(smem_ptr);
uint64_t state = 0;
asm volatile(
"{\n\t"
"mbarrier.arrive.shared.b64 %1, [%0];\n\t"
"mbarrier.arrive.shared.b64 _, [%0];\n\t"
"}"
:
: "r"(smem_addr), "l"(state));
: "r"(smem_addr));
#elif defined(__CUDA_ARCH__)
asm volatile ("brkpt;\n" ::);
#endif