Skip to content
GitLab
Menu
Projects
Groups
Snippets
Loading...
Help
Help
Support
Community forum
Keyboard shortcuts
?
Submit feedback
Contribute to GitLab
Sign in
Toggle navigation
Menu
Open sidebar
OpenDAS
FlashMLA
Commits
c353b35b
Commit
c353b35b
authored
Feb 27, 2026
by
zhanghj2
Browse files
恢复支持旧接口
parent
c566af36
Changes
27
Expand all
Hide whitespace changes
Inline
Side-by-side
Showing
7 changed files
with
1351 additions
and
155 deletions
+1351
-155
flash_mla/flash_mla_interface.py
flash_mla/flash_mla_interface.py
+480
-116
setup.py
setup.py
+10
-1
tests/test_flash_mla_fp8.py
tests/test_flash_mla_fp8.py
+11
-31
tests/test_flash_mla_qkvfp8.py
tests/test_flash_mla_qkvfp8.py
+13
-7
tests/test_flash_mla_qkvfp8_with_cat.py
tests/test_flash_mla_qkvfp8_with_cat.py
+274
-0
tests/test_flash_mla_with_q_concat.py
tests/test_flash_mla_with_q_concat.py
+319
-0
tests/test_flash_mla_with_q_concat_fp8.py
tests/test_flash_mla_with_q_concat_fp8.py
+244
-0
No files found.
flash_mla/flash_mla_interface.py
View file @
c353b35b
This diff is collapsed.
Click to expand it.
setup.py
View file @
c353b35b
...
@@ -85,6 +85,12 @@ ext_modules.append(
...
@@ -85,6 +85,12 @@ ext_modules.append(
"csrc/gfx93/prefill/sparse/instantiations/phase1_k576.cu"
,
"csrc/gfx93/prefill/sparse/instantiations/phase1_k576.cu"
,
"csrc/gfx93/prefill/sparse/instantiations/phase1_k576_topklen.cu"
,
"csrc/gfx93/prefill/sparse/instantiations/phase1_k576_topklen.cu"
,
"csrc/extension/flash_fwd_mla_bf16_gfx936.cu"
,
"csrc/extension/flash_fwd_mla_fp16_gfx936.cu"
,
"csrc/extension/flash_fwd_mla_fp8_gfx938.cu"
,
"csrc/extension/flash_fwd_mla_fp8_qbf16_gfx938.cu"
,
"csrc/extension/flash_fwd_mla_metadata.cu"
,
],
],
extra_compile_args
=
{
extra_compile_args
=
{
"cxx"
:
cxx_args
+
get_features_args
(),
"cxx"
:
cxx_args
+
get_features_args
(),
...
@@ -98,7 +104,9 @@ ext_modules.append(
...
@@ -98,7 +104,9 @@ ext_modules.append(
"-Rpass-analysis=kernel-resource-usage"
,
"-Rpass-analysis=kernel-resource-usage"
,
"-DDCU_ASM"
,
"-DDCU_ASM"
,
"--save-temps"
,
"--save-temps"
,
"-w"
"-w"
,
"-mllvm"
,
"-enable-num-vgprs-512=true"
,
]
+
get_features_args
()
+
get_arch_flags
()
]
+
get_features_args
()
+
get_arch_flags
()
},
},
include_dirs
=
[
include_dirs
=
[
...
@@ -134,5 +142,6 @@ setup(
...
@@ -134,5 +142,6 @@ setup(
version
=
get_version
(
ROCM_HOME
),
version
=
get_version
(
ROCM_HOME
),
packages
=
find_packages
(
include
=
[
'flash_mla'
]),
packages
=
find_packages
(
include
=
[
'flash_mla'
]),
ext_modules
=
ext_modules
,
ext_modules
=
ext_modules
,
package_data
=
{
"flash_mla"
:[
"asm/*.co"
]},
cmdclass
=
{
"build_ext"
:
BuildExtension
},
cmdclass
=
{
"build_ext"
:
BuildExtension
},
)
)
tests/test_flash_mla_
kv
fp8.py
→
tests/test_flash_mla_fp8.py
View file @
c353b35b
This diff is collapsed.
Click to expand it.
tests/test_flash_mla_qkvfp8.py
View file @
c353b35b
This diff is collapsed.
Click to expand it.
tests/test_flash_mla_qkvfp8_with_cat.py
0 → 100644
View file @
c353b35b
This diff is collapsed.
Click to expand it.
tests/test_flash_mla_with_q_concat.py
0 → 100644
View file @
c353b35b
This diff is collapsed.
Click to expand it.
tests/test_flash_mla_with_q_concat_fp8.py
0 → 100644
View file @
c353b35b
This diff is collapsed.
Click to expand it.
Prev
1
2
Next
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