Skip to content
GitLab
Menu
Projects
Groups
Snippets
Loading...
Help
Help
Support
Community forum
Keyboard shortcuts
?
Submit feedback
Contribute to GitLab
Sign in / Register
Toggle navigation
Menu
Open sidebar
gaoqiong
composable_kernel_ROCM
Commits
fe728dc5
Commit
fe728dc5
authored
Apr 18, 2024
by
illsilin
Browse files
disable instances with blocksize=256 in attention examples
parent
9b69d57f
Changes
2
Hide whitespace changes
Inline
Side-by-side
Showing
2 changed files
with
6 additions
and
6 deletions
+6
-6
example/32_batched_gemm_scale_softmax_gemm/cross_attention_forward_wmma_fp16.cpp
..._scale_softmax_gemm/cross_attention_forward_wmma_fp16.cpp
+3
-3
example/32_batched_gemm_scale_softmax_gemm/self_attention_forward_wmma_fp16.cpp
...m_scale_softmax_gemm/self_attention_forward_wmma_fp16.cpp
+3
-3
No files found.
example/32_batched_gemm_scale_softmax_gemm/cross_attention_forward_wmma_fp16.cpp
View file @
fe728dc5
...
...
@@ -71,7 +71,7 @@ static constexpr auto TensorSpecC = ck::tensor_operation::device::TensorSpecial
#define CK_MHA_USE_WAVE_1
#define CK_MHA_USE_WAVE_2
#define CK_MHA_USE_WAVE_4
#define CK_MHA_USE_WAVE_8
//
#define CK_MHA_USE_WAVE_8
using
DeviceMHAFactory
=
std
::
tuple
<
#ifdef CK_MHA_USE_WAVE_1
...
...
@@ -277,10 +277,10 @@ using DeviceMHAFactory =
S
<
2
,
8
,
8
>
,
S
<
0
,
2
,
1
>
,
S
<
0
,
2
,
1
>
,
1
,
2
,
1
,
false
,
// CShuffleBlockTransfer MN
1
,
1
,
S
<
1
,
64
,
1
,
2
>
,
8
,
MaskingSpec
>
,
MaskingSpec
>
#endif
#ifdef CK_MHA_USE_WAVE_8
ck
::
tensor_operation
::
device
::
DeviceBatchedGemmSoftmaxGemmPermute_Wmma_CShuffle
<
,
ck
::
tensor_operation
::
device
::
DeviceBatchedGemmSoftmaxGemmPermute_Wmma_CShuffle
<
NumDimG
,
NumDimM
,
NumDimN
,
NumDimK
,
NumDimO
,
ADataType
,
B0DataType
,
B1DataType
,
CDataType
,
Acc0BiasDataType
,
Acc0DataType
,
Acc1BiasDataType
,
Acc1DataType
,
CShuffleDataType
,
AElementOp
,
B0ElementOp
,
Acc0ElementOp
,
B1ElementOp
,
CElementOp
,
...
...
example/32_batched_gemm_scale_softmax_gemm/self_attention_forward_wmma_fp16.cpp
View file @
fe728dc5
...
...
@@ -71,7 +71,7 @@ static constexpr auto TensorSpecC = ck::tensor_operation::device::TensorSpecial
#define CK_MHA_USE_WAVE_1
#define CK_MHA_USE_WAVE_2
#define CK_MHA_USE_WAVE_4
#define CK_MHA_USE_WAVE_8
//
#define CK_MHA_USE_WAVE_8
using
DeviceMHAFactory
=
std
::
tuple
<
#ifdef CK_MHA_USE_WAVE_1
...
...
@@ -277,10 +277,10 @@ using DeviceMHAFactory =
S
<
2
,
8
,
8
>
,
S
<
0
,
2
,
1
>
,
S
<
0
,
2
,
1
>
,
1
,
2
,
1
,
false
,
// CShuffleBlockTransfer MN
1
,
1
,
S
<
1
,
64
,
1
,
2
>
,
8
,
MaskingSpec
>
,
MaskingSpec
>
#endif
#ifdef CK_MHA_USE_WAVE_8
ck
::
tensor_operation
::
device
::
DeviceBatchedGemmSoftmaxGemmPermute_Wmma_CShuffle
<
,
ck
::
tensor_operation
::
device
::
DeviceBatchedGemmSoftmaxGemmPermute_Wmma_CShuffle
<
NumDimG
,
NumDimM
,
NumDimN
,
NumDimK
,
NumDimO
,
ADataType
,
B0DataType
,
B1DataType
,
CDataType
,
Acc0BiasDataType
,
Acc0DataType
,
Acc1BiasDataType
,
Acc1DataType
,
CShuffleDataType
,
AElementOp
,
B0ElementOp
,
Acc0ElementOp
,
B1ElementOp
,
CElementOp
,
...
...
Write
Preview
Markdown
is supported
0%
Try again
or
attach a new file
.
Attach a file
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Cancel
Please
register
or
sign in
to comment