wu-arch/kernels
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Compare commits

merge into: wu-arch:ae
Branches Tags
wu-arch:ae wu-arch:ae-flash-virgo wu-arch:ae-flash-ampere wu-arch:ae-hopper wu-arch:ae-volta wu-arch:ae-ampere
...
pull from: wu-arch:ae-volta
Branches Tags
wu-arch:ae wu-arch:ae-flash-virgo wu-arch:ae-flash-ampere wu-arch:ae-hopper wu-arch:ae-volta wu-arch:ae-ampere

13 Commits
ae ... ae-volta

Author SHA1 Message Date
Virgo-AE Eval
326141b11f Merge branch 'ae' into ae-volta 2025-02-07 14:52:25 -08:00
Richard Yan
d893780594 Merge branch 'ae' into ae-volta 2025-01-31 03:53:26 -08:00
Richard Yan
9b7c22a7e9 Merge branch 'ae' into ae-volta 2025-01-30 23:41:16 -08:00
Richard Yan
b1ebabef26 Merge branch 'ae' into ae-volta 2025-01-30 15:35:22 -08:00
Hansung Kim
9f524538a4 Merge branch 'ae' into ae-volta 2025-01-30 13:24:46 -08:00
Hansung Kim
51ebe18ebb Merge remote-tracking branch 'origin/ae-volta' into ae-volta 2025-01-30 01:48:25 -08:00
Hansung Kim
7b0a95034b Merge branch 'ae' into ae-volta 2025-01-30 01:48:05 -08:00
Richard Yan
c240069147 Merge branch 'ae' into ae-volta 2025-01-30 01:35:35 -08:00
Hansung Kim
d86c33acf3 Merge branch 'ae' into ae-volta 2025-01-30 01:05:27 -08:00
Hansung Kim
b49e8a293c Merge branch 'ae' into ae-volta 2025-01-30 00:49:19 -08:00
Hansung Kim
19731b8e2f Merge branch 'ae' into ae-volta 2025-01-30 00:35:00 -08:00
Richard Yan
afc69507a3 Merge branch 'ae' into ae-volta 2025-01-29 23:31:34 -08:00
Richard Yan
6e279c905f volta change 2025-01-29 22:16:39 -08:00
1 changed files with 1 additions and 1 deletions
Expand all files Collapse all files

2
kernels/sgemm_tcore/sgemm_impl.hpp
View File

@@ -110,7 +110,7 @@ static_assert(WMITER * WNITER * TCM * TCN * NUM_WARPS * CORES_PER_CLUSTER ==
// result matrix will be stored in a swizzled form in the global memory. // result matrix will be stored in a swizzled form in the global memory.
#define WMMA_STORE_FAST 0 #define WMMA_STORE_FAST 0
#define GEMMINI_DMA 1 #define GEMMINI_DMA 0
#define GEMMINI_DMA_FAST 1 #define GEMMINI_DMA_FAST 1
#if SMEM_SIZE == 0x4000 #if SMEM_SIZE == 0x4000
#define SMEM_ADDR_Q0 ((float * const) 0xff000000) #define SMEM_ADDR_Q0 ((float * const) 0xff000000)