Skip to content

GitLab

Commit 514cee8a authored by letaoqin's avatar letaoqin
Browse files

bias parameters update to other example

parent d84b00c7
Expand all Show whitespace changes
Inline Side-by-side
Showing with 48 additions and 36 deletions
example/32_batched_gemm_scale_softmax_gemm/batched_multihead_attention_forward_v2.cpp
View file @ 514cee8a
...@@ -135,6 +135,7 @@ using DeviceGemmInstance = ...@@ -135,6 +135,7 @@ using DeviceGemmInstance =
8, 8,
8, 8,
true, true,
4,
S<16, 16, 1>, // B1BlockTransfer S<16, 16, 1>, // B1BlockTransfer
S<0, 2, 1>, S<0, 2, 1>,
S<0, 2, 1>, S<0, 2, 1>,
...@@ -146,6 +147,7 @@ using DeviceGemmInstance = ...@@ -146,6 +147,7 @@ using DeviceGemmInstance =
1, // CShuffleNXdlPerWavePerShuffle 1, // CShuffleNXdlPerWavePerShuffle
S<1, 64, 1, 4>, // CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock S<1, 64, 1, 4>, // CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock
8, // CShuffleBlockTransferScalarPerVector_NPerBlock 8, // CShuffleBlockTransferScalarPerVector_NPerBlock
4,
MaskingSpec, // MaskingSpecialization MaskingSpec, // MaskingSpecialization
Deterministic>; Deterministic>;
#elif(DIM <= 64) #elif(DIM <= 64)
...@@ -206,6 +208,7 @@ using DeviceGemmInstance = ...@@ -206,6 +208,7 @@ using DeviceGemmInstance =
8, 8,
8, 8,
true, true,
4,
S<16, 16, 1>, // B1BlockTransfer S<16, 16, 1>, // B1BlockTransfer
S<0, 2, 1>, S<0, 2, 1>,
S<0, 2, 1>, S<0, 2, 1>,
...@@ -217,6 +220,7 @@ using DeviceGemmInstance = ...@@ -217,6 +220,7 @@ using DeviceGemmInstance =
2, // CShuffleNXdlPerWavePerShuffle 2, // CShuffleNXdlPerWavePerShuffle
S<1, 32, 1, 8>, // CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock S<1, 32, 1, 8>, // CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock
8, // CShuffleBlockTransferScalarPerVector_NPerBlock 8, // CShuffleBlockTransferScalarPerVector_NPerBlock
4,
MaskingSpec, // MaskingSpecialization MaskingSpec, // MaskingSpecialization
Deterministic>; Deterministic>;
#elif(DIM <= 128) #elif(DIM <= 128)
...@@ -277,6 +281,7 @@ using DeviceGemmInstance = ...@@ -277,6 +281,7 @@ using DeviceGemmInstance =
8, 8,
8, 8,
true, true,
4,
S<8, 32, 1>, // B1BlockTransfer S<8, 32, 1>, // B1BlockTransfer
S<0, 2, 1>, S<0, 2, 1>,
S<0, 2, 1>, S<0, 2, 1>,
...@@ -288,6 +293,7 @@ using DeviceGemmInstance = ...@@ -288,6 +293,7 @@ using DeviceGemmInstance =
2, // CShuffleNXdlPerWavePerShuffle 2, // CShuffleNXdlPerWavePerShuffle
S<1, 32, 1, 8>, // CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock S<1, 32, 1, 8>, // CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock
8, // CShuffleBlockTransferScalarPerVector_NPerBlock 8, // CShuffleBlockTransferScalarPerVector_NPerBlock
4,
MaskingSpec, // MaskingSpecialization MaskingSpec, // MaskingSpecialization
Deterministic>; Deterministic>;
#endif #endif
......
example/32_batched_gemm_scale_softmax_gemm/grouped_multihead_attention_forward_v2.cpp
View file @ 514cee8a
...@@ -135,6 +135,7 @@ using DeviceGemmInstance = ...@@ -135,6 +135,7 @@ using DeviceGemmInstance =
8, 8,
8, 8,
true, true,
1,
S<16, 16, 1>, // B1BlockTransfer S<16, 16, 1>, // B1BlockTransfer
S<0, 2, 1>, S<0, 2, 1>,
S<0, 2, 1>, S<0, 2, 1>,
...@@ -146,6 +147,7 @@ using DeviceGemmInstance = ...@@ -146,6 +147,7 @@ using DeviceGemmInstance =
1, // CShuffleNXdlPerWavePerShuffle 1, // CShuffleNXdlPerWavePerShuffle
S<1, 64, 1, 4>, // CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock S<1, 64, 1, 4>, // CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock
8, // CShuffleBlockTransferScalarPerVector_NPerBlock 8, // CShuffleBlockTransferScalarPerVector_NPerBlock
1,
MaskingSpec, // MaskingSpecialization MaskingSpec, // MaskingSpecialization
Deterministic>; Deterministic>;
#elif(DIM <= 64) #elif(DIM <= 64)
...@@ -206,6 +208,7 @@ using DeviceGemmInstance = ...@@ -206,6 +208,7 @@ using DeviceGemmInstance =
8, 8,
8, 8,
true, true,
1,
S<16, 16, 1>, // B1BlockTransfer S<16, 16, 1>, // B1BlockTransfer
S<0, 2, 1>, S<0, 2, 1>,
S<0, 2, 1>, S<0, 2, 1>,
...@@ -217,6 +220,7 @@ using DeviceGemmInstance = ...@@ -217,6 +220,7 @@ using DeviceGemmInstance =
2, // CShuffleNXdlPerWavePerShuffle 2, // CShuffleNXdlPerWavePerShuffle
S<1, 32, 1, 8>, // CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock S<1, 32, 1, 8>, // CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock
8, // CShuffleBlockTransferScalarPerVector_NPerBlock 8, // CShuffleBlockTransferScalarPerVector_NPerBlock
1,
MaskingSpec, // MaskingSpecialization MaskingSpec, // MaskingSpecialization
Deterministic>; Deterministic>;
#elif(DIM <= 128) #elif(DIM <= 128)
...@@ -277,6 +281,7 @@ using DeviceGemmInstance = ...@@ -277,6 +281,7 @@ using DeviceGemmInstance =
8, 8,
8, 8,
true, true,
1,
S<8, 32, 1>, // B1BlockTransfer S<8, 32, 1>, // B1BlockTransfer
S<0, 2, 1>, S<0, 2, 1>,
S<0, 2, 1>, S<0, 2, 1>,
...@@ -288,6 +293,7 @@ using DeviceGemmInstance = ...@@ -288,6 +293,7 @@ using DeviceGemmInstance =
2, // CShuffleNXdlPerWavePerShuffle 2, // CShuffleNXdlPerWavePerShuffle
S<1, 32, 1, 8>, // CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock S<1, 32, 1, 8>, // CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock
8, // CShuffleBlockTransferScalarPerVector_NPerBlock 8, // CShuffleBlockTransferScalarPerVector_NPerBlock
1,
MaskingSpec, // MaskingSpecialization MaskingSpec, // MaskingSpecialization
Deterministic>; Deterministic>;
#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