add pytest support for tutorial gemm (#2826)
* add pytest support for tutorial gemm * add license
This commit is contained in:
@@ -141,15 +141,15 @@ def kernel(
|
||||
# (bM, bN)
|
||||
gC = cute.local_tile(mC_mnl, mma_tiler_mnk, mma_coord_mnk, proj=(1, 1, None))
|
||||
thr_mma = tiled_mma.get_slice(0)
|
||||
# (MMA, MMA_M, MMA_K)
|
||||
# (MMA, MMA_M, MMA_K, RestK)
|
||||
tCgA = thr_mma.partition_A(gA)
|
||||
# (MMA, MMA_N, MMA_K)
|
||||
# (MMA, MMA_N, MMA_K, RestK)
|
||||
tCgB = thr_mma.partition_B(gB)
|
||||
# (MMA, MMA_M, MMA_N)
|
||||
tCgC = thr_mma.partition_C(gC)
|
||||
# (MMA, MMA_M, MMA_K)
|
||||
# (MMA, MMA_M, MMA_K, STAGE)
|
||||
tCrA = tiled_mma.make_fragment_A(sA)
|
||||
# (MMA, MMA_N, MMA_K)
|
||||
# (MMA, MMA_N, MMA_K, STAGE)
|
||||
tCrB = tiled_mma.make_fragment_B(sB)
|
||||
# (MMA, MMA_M, MMA_N)
|
||||
acc_shape = tiled_mma.partition_shape_C(mma_tiler_mnk[:2])
|
||||
|
||||
@@ -174,15 +174,15 @@ def kernel(
|
||||
# (bM, bN)
|
||||
gC = cute.local_tile(mC_mnl, mma_tiler_mnk, mma_coord_mnk, proj=(1, 1, None))
|
||||
thr_mma = tiled_mma.get_slice(mma_coord_vmnk[0])
|
||||
# (MMA, MMA_M, MMA_K)
|
||||
# (MMA, MMA_M, MMA_K, RestK)
|
||||
tCgA = thr_mma.partition_A(gA)
|
||||
# (MMA, MMA_N, MMA_K)
|
||||
# (MMA, MMA_N, MMA_K, RestK)
|
||||
tCgB = thr_mma.partition_B(gB)
|
||||
# (MMA, MMA_M, MMA_N)
|
||||
tCgC = thr_mma.partition_C(gC)
|
||||
# (MMA, MMA_M, MMA_K)
|
||||
# (MMA, MMA_M, MMA_K, STAGE)
|
||||
tCrA = tiled_mma.make_fragment_A(sA)
|
||||
# (MMA, MMA_N, MMA_K)
|
||||
# (MMA, MMA_N, MMA_K, STAGE)
|
||||
tCrB = tiled_mma.make_fragment_B(sB)
|
||||
# (MMA, MMA_M, MMA_N)
|
||||
acc_shape = tiled_mma.partition_shape_C(mma_tiler_mnk[:2])
|
||||
|
||||
Reference in New Issue
Block a user