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
Commits
2e1f2c37
"docs/source/api/pipelines/stable_diffusion/overview.mdx" did not exist on "0eb507f2af991b1f0b6c2ede5b20a994999e85d3"
Commit
2e1f2c37
authored
Mar 02, 2023
by
ltqin
Browse files
fix datatype
parent
32b03f33
Changes
2
Show whitespace changes
Inline
Side-by-side
Showing
2 changed files
with
3 additions
and
3 deletions
+3
-3
include/ck/tensor_operation/gpu/grid/gridwise_batched_multihead_attention_backward_xdl_cshuffle_pt1.hpp
...batched_multihead_attention_backward_xdl_cshuffle_pt1.hpp
+2
-2
include/ck/tensor_operation/gpu/grid/gridwise_batched_multihead_attention_backward_xdl_cshuffle_v2.hpp
..._batched_multihead_attention_backward_xdl_cshuffle_v2.hpp
+1
-1
No files found.
include/ck/tensor_operation/gpu/grid/gridwise_batched_multihead_attention_backward_xdl_cshuffle_pt1.hpp
View file @
2e1f2c37
include/ck/tensor_operation/gpu/grid/gridwise_batched_multihead_attention_backward_xdl_cshuffle_v2.hpp
View file @
2e1f2c37
...
@@ -908,7 +908,7 @@ struct GridwiseBatchedMultiheadAttentionBackward_Xdl_CShuffle_V2
...
@@ -908,7 +908,7 @@ struct GridwiseBatchedMultiheadAttentionBackward_Xdl_CShuffle_V2
template
<
index_t
BlockSize_
,
index_t
BlockSliceLength_M_
,
index_t
BlockSliceLength_O_
>
template
<
index_t
BlockSize_
,
index_t
BlockSliceLength_M_
,
index_t
BlockSliceLength_O_
>
struct
YDotYGrad_M_O_
struct
YDotYGrad_M_O_
{
{
static
constexpr
index_t
SrcScalarPerVector
=
16
/
sizeof
(
FloatGemmAcc
);
static
constexpr
index_t
SrcScalarPerVector
=
16
/
sizeof
(
DataType
);
static
constexpr
auto
ThreadClusterLength_O
=
static
constexpr
auto
ThreadClusterLength_O
=
Number
<
BlockSliceLength_O_
/
SrcScalarPerVector
>
{};
Number
<
BlockSliceLength_O_
/
SrcScalarPerVector
>
{};
static
constexpr
auto
ThreadClusterLength_M
=
Number
<
BlockSize_
/
ThreadClusterLength_O
>
{};
static
constexpr
auto
ThreadClusterLength_M
=
Number
<
BlockSize_
/
ThreadClusterLength_O
>
{};
...
...
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