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
3e411ef0
Commit
3e411ef0
authored
Jan 31, 2025
by
Qianfeng Zhang
Browse files
Use MakeQDramTileDistribution for q_dram_window
parent
ceea2722
Changes
1
Hide whitespace changes
Inline
Side-by-side
Showing
1 changed file
with
6 additions
and
5 deletions
+6
-5
include/ck_tile/ops/fmha/pipeline/block_fmha_pipeline_qr_ks_vs_async.hpp
.../ops/fmha/pipeline/block_fmha_pipeline_qr_ks_vs_async.hpp
+6
-5
No files found.
include/ck_tile/ops/fmha/pipeline/block_fmha_pipeline_qr_ks_vs_async.hpp
View file @
3e411ef0
...
...
@@ -175,11 +175,12 @@ struct BlockFmhaPipelineQRKSVSAsync
static_assert
(
NumKLdsBuffers
>=
2
);
static_assert
(
NumVLdsBuffers
>=
2
);
auto
q_dram_window
=
make_tile_window
(
q_dram_block_window_tmp
.
get_bottom_tensor_view
(),
q_dram_block_window_tmp
.
get_window_lengths
(),
q_dram_block_window_tmp
.
get_window_origin
(),
Policy
::
template
MakeQRegTileDistribution
<
Problem
>());
auto
original_q
=
load_tile
(
q_dram_window
);
auto
q_dram_window
=
make_tile_window
(
q_dram_block_window_tmp
.
get_bottom_tensor_view
(),
q_dram_block_window_tmp
.
get_window_lengths
(),
q_dram_block_window_tmp
.
get_window_origin
(),
Policy
::
template
MakeQDramTileDistribution
<
Problem
>());
auto
original_q
=
load_tile
(
q_dram_window
);
__builtin_amdgcn_sched_barrier
(
0
);
...
...
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