Skip to content
Open
Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
188 commits
Select commit Hold shift + click to select a range
085b010
[BUILD] Change hatchet as an extra dependency (#4138) (#4158)
atalman Jun 18, 2024
45b0f6b
Use dictionary unpacking instead of merge operator for python 3.8 com…
amjames Jun 18, 2024
3d1da66
[AMD] Move MFMA shortcut check to not compute scratch buffer shape if…
amjames Jun 18, 2024
dcc22b8
Remove redundant options from passes (#4015) (#4162)
amjames Jun 18, 2024
5d77ae2
[AMD] Add more math functions in libdevice (#4086) (#4163)
amjames Jun 18, 2024
dedb7bd
[BACKEND] Prevent for/yield argument number drift (#4097) (#4164)
amjames Jun 18, 2024
21eae95
[RELEASE] [AMD] Additional AMD cherry-picks (#4175)
jataylo Jun 20, 2024
107465b
[RELEASE][BACKEND] Fix getThreadsPerWarp for MD sliced encoding (#406…
htyu Jun 28, 2024
a378f68
[CI] update wheels workflow
ptillet Jul 9, 2024
72b422d
[CI] exclude python 3.7 from wheels workflow
ptillet Jul 9, 2024
464c6b3
[AMD] NFC:Adding AMD GPUs to official support in release branch (#4285)
jerryyin Jul 13, 2024
5cbb516
[VERSION] Bumped to 3.0.0-post1 (#4340)
atalman Jul 17, 2024
91f24d8
[Runtime] Dynamically load cuTensorMapEncodeTiled (#4330) (#4339)
atalman Jul 17, 2024
b514024
Repack wheels with build-number 1 (#4354)
malfet Jul 18, 2024
be5e5f1
Remove quotes around wheel in wheel unpack
atalman Jul 18, 2024
55a4ab0
Use same bash as other steps in Append build number. Install pip whee…
atalman Jul 19, 2024
757b6a6
[AMD] Cherry-pick commits from mainline to support Flex attention on …
jerrymannil Aug 26, 2024
5fe38ff
[Release 3.1.0] Bump version to 3.1.0 and remove temporary change (#4…
atalman Sep 10, 2024
23fefa6
Cherry Pick of #4311 to release/3.1.x branch (#4705)
drisspg Sep 13, 2024
fc879bd
Cherry Pick of #4247 to release/3.1.x branch (#4706)
drisspg Sep 13, 2024
566b63c
[RELEASE] Cherry-pick use of device-agnostic `DeviceInterface` in `do…
galexite Sep 16, 2024
cf34004
AMD requested cherry-picks for release/3.1.x (#4794)
bertmaher Sep 24, 2024
819b371
Add string representation for AttrsDescriptor (#4888)
alexbaden Oct 29, 2024
d47fa58
Update version to 3.2.0
bertmaher Oct 30, 2024
97f6cd0
[Backend] Fix predicates for device assert inside reduction/scan regi…
davidberard98 Nov 5, 2024
9c689f5
[BACKEND] Fix asserts in 2d scan and add assert mode to layout tests …
peterbell10 Nov 5, 2024
b5ceca3
Add back barrier after asserts (#5043)
ThomasRaoux Nov 2, 2024
90d44e4
[FRONTEND] Fix handling of `from m import x as y` in CodeGenerator (#…
davidberard98 Nov 6, 2024
c2143cb
[BACKEND] Make ExternElementwise op implement ConditionallySpeculatab…
davidberard98 Nov 6, 2024
601aec2
[RUNTIME] Pass full kwargs to Autotuner hooks instead of positional a…
aakhundov Nov 7, 2024
24c0fe4
[BACKEND] Fix ProgramPoint passing in AxisInfoAnalysis (#5181)
aakhundov Nov 18, 2024
3e00b0e
[RUNTIME] Add flags for detecting user-defined Autotuner hooks (#5092)
aakhundov Nov 9, 2024
35c6c7c
Revert "[BACKEND] Optimize code generation for load with other arg (#…
davidberard98 Nov 20, 2024
dbc771e
[Release Only Changes] Remove git commit hash in wheel name (#5405)
atalman Dec 11, 2024
8af9311
[release-only] Add manylinux2014_x86_64 for PyTorch release 2.6 (#5414)
atalman Dec 13, 2024
e74f027
[BACKEND] Fix and document logic for creating warp shapes in MMAv3 (#…
bertmaher Dec 17, 2024
aba0fbf
[release/3.2.x] [CHERRY PICK] Add gfx950 target definition (#5452)
jataylo Dec 19, 2024
64b80f0
[release/3.2.x] [CHERRY PICK] [AMD] Fix issue with rank=1 in tryFitCv…
jataylo Dec 23, 2024
ebffad5
Automatic Warp Specialization Optimization (#5622)
htyu Jan 15, 2025
6771065
[cherrypick release/3.2][BACKEND] Fix accumulator init optimization f…
bertmaher Jan 15, 2025
e101b08
[3.2 cherry pick] Ensure device context before launching kernel (#373…
bertmaher Jan 21, 2025
9641643
Release Only - Enable pypi promotion for 3.2.x release (#5618)
atalman Jan 22, 2025
9336065
[INIT] Init flagtree (#1)
zhzhcookie Mar 25, 2025
0f74943
[CI/CD] Init flagtree workflow (#2)
zhzhcookie Mar 26, 2025
f41876b
[DOC] [BUILD] Update build deps download addr (#3)
zhzhcookie Mar 26, 2025
3548338
[DOC] Update readme and release notes (#4)
zhzhcookie Mar 26, 2025
23d2857
[BUILD] Fix ext_sourcedir in editable_wheel mode, and update require…
zhzhcookie Mar 26, 2025
b382dd8
[CI] Fix build and test workflow (#6)
zhzhcookie Mar 26, 2025
a27316f
Revert "AMD requested cherry-picks for release/3.1.x (#4794)"
zhzhcookie Mar 31, 2025
a69557d
Revert "[RELEASE] Cherry-pick use of device-agnostic `DeviceInterface…
zhzhcookie Mar 31, 2025
fc9e223
Revert "Cherry Pick of #4247 to release/3.1.x branch (#4706)"
zhzhcookie Mar 31, 2025
7a9e367
Revert "Cherry Pick of #4311 to release/3.1.x branch (#4705)"
zhzhcookie Mar 31, 2025
1c3ee97
Revert "[Release 3.1.0] Bump version to 3.1.0 and remove temporary ch…
zhzhcookie Mar 31, 2025
af32f8a
Revert "[AMD] Cherry-pick commits from mainline to support Flex atten…
zhzhcookie Mar 31, 2025
7b5fffc
Revert "Repack wheels with build-number 1 (#4354)"
zhzhcookie Mar 31, 2025
6ea9e5f
Revert "[Runtime] Dynamically load cuTensorMapEncodeTiled (#4330) (#4…
zhzhcookie Mar 31, 2025
13b1932
Revert "[VERSION] Bumped to 3.0.0-post1 (#4340)"
zhzhcookie Mar 31, 2025
13a139b
Revert "[RELEASE][BACKEND] Fix getThreadsPerWarp for MD sliced encodi…
zhzhcookie Mar 31, 2025
d7e6ce2
Revert "[RELEASE] [AMD] Additional AMD cherry-picks (#4175)"
zhzhcookie Mar 31, 2025
e0cd80d
Revert "[BACKEND] Prevent for/yield argument number drift (#4097) (#4…
zhzhcookie Mar 31, 2025
a0ac7d8
Revert "[AMD] Add more math functions in libdevice (#4086) (#4163)"
zhzhcookie Mar 31, 2025
efea94b
Revert "Remove redundant options from passes (#4015) (#4162)"
zhzhcookie Mar 31, 2025
d5c6092
Revert "[AMD] Move MFMA shortcut check to not compute scratch buffer …
zhzhcookie Mar 31, 2025
3e3afef
Revert "Use dictionary unpacking instead of merge operator for python…
zhzhcookie Mar 31, 2025
e693ac5
Revert "[BUILD] Change hatchet as an extra dependency (#4138) (#4158)"
zhzhcookie Mar 31, 2025
b65638a
Merge remote-tracking branch 'upstream/release/3.2.x' into triton_v3.2.x
zhzhcookie May 15, 2025
72053b3
[BUILD] Fix merge Triton 3.2.x, disable triton_shared temporarily
zhzhcookie May 15, 2025
05be61b
[BACKEND] Add ascend backend (#8)
zhzhcookie Jun 4, 2025
3a72726
[BUILD] Update build ascend backend (#10)
zhzhcookie Jun 6, 2025
97e653a
[DOC] Update readme (#14)
zhzhcookie Jun 6, 2025
4d19930
[install & compile refactor] refactor setup (#19)
Galaxy1458 Jun 16, 2025
b3e9e89
[DOC] Update build command in readme (#33)
zhzhcookie Jun 25, 2025
a72467f
[Ascend] Fix ascend patch (#53)
Galaxy1458 Jul 30, 2025
dba633f
[BACKEND] Support cambricon (#54) move to triton_v3.2.x (#56)
zhzhcookie Jul 30, 2025
0ea2396
[3.2 build infra]modify build infra (#57)
Galaxy1458 Aug 1, 2025
7f0bbc1
[ascend]fix bug (#63)
Galaxy1458 Aug 20, 2025
594f5d6
[Ascend]Ascend bug fix (#71)
Galaxy1458 Aug 26, 2025
eb0368f
[CI/CD] Update ascend workflow (#73)
zhzhcookie Aug 27, 2025
516d783
[BACKEND] Update ascend dockerfile (#83)
zhzhcookie Sep 8, 2025
9eb36a4
Update ascend.py (#93)
Galaxy1458 Sep 12, 2025
ca4c0f0
upcase as to solve warnings (#95)
sgjzfzzf Sep 15, 2025
77050a9
[Common]Add UnifiedHardware (#39) (#96)
sgjzfzzf Sep 16, 2025
0605d37
[Ascend ]Update new repo for ascend-third-party (#102)
Galaxy1458 Sep 19, 2025
4a83a0b
[BUILD] Update SDK from github to ksyuncs (#104)
zhzhcookie Sep 22, 2025
dd0bdee
[Build] Cherry-pick build_infra_modified from v3.1(main) to v3.2 (#111)
Galaxy1458 Sep 25, 2025
b4b84c7
alter_commit_and_fix_bugs (#120)
Galaxy1458 Oct 10, 2025
2ba8190
Update submodule URL and commit ID for ascendnpu-ir (#128)
Galaxy1458 Oct 21, 2025
15f4acc
Fix path reference in git clone command (#130)
Galaxy1458 Oct 22, 2025
816b53c
[BUILD] Offline Build Scripts for Triton-v3.2.x (#133)
liuyunqi20 Oct 28, 2025
1c6cc27
[FixBug]Fixed the bug where the submodule npu-ir could not be pulled …
SabYic Oct 29, 2025
18dd76f
[CI/CD] Update workflow repo addr to flagos-ai on branch triton_v3.2.…
zhzhcookie Nov 13, 2025
c001cc4
[BACKEND] [BUILD] Enhance ascend submodule commit pull action (#149)
Galaxy1458 Nov 17, 2025
999b35a
[BACKEND] update cambricon (#161)
chenmiao1919 Dec 1, 2025
cc63bc4
[BUILD] Update setup info for pypi (#163)
zhzhcookie Dec 2, 2025
43c12a8
[DOC] Update aipu llvm readme (#180)
zhzhcookie Dec 10, 2025
eefc0a3
[BACKEND] Fix autotuner on cambricon backend (#189)
chenmiao1919 Dec 12, 2025
c91dd75
[BACKEND] Fix autotuner on cambricon backend_ (#194)
chenmiao1919 Dec 16, 2025
8355aaf
[BUILD] ascend with flir and local ascend dir (#135)
zhzhcookie Dec 22, 2025
4a4af52
fix_asecnd_setup_bug (#224)
Galaxy1458 Dec 24, 2025
70d3aac
[BUILD] Invert flag conditions for third-party flags for triton_v3.2.…
Galaxy1458 Dec 25, 2025
aa92d8b
[CI/CD] Update doc workflow (#232)
zhzhcookie Dec 25, 2025
0e006dd
Open [CI/CD] Update workflow to check doc changed for triton 3.2 (#240)
zhzhcookie Dec 26, 2025
3fb2b77
[BACKEND] Update ascend backend to 5b0b2fce on triton-ascend (#219)
ph0375 Jan 7, 2026
45f2eb1
[FLIR] Update flir npu path (#280)
zhzhcookie Jan 8, 2026
a51b967
[Release] Update flagtree version to v0.4.0 on triton_v3.2.x branch (…
zhzhcookie Jan 8, 2026
4594473
[Decoupling] Fix new py files which ascend added (#286)
zhzhcookie Jan 12, 2026
e6209f2
[BUILD] Change build option to LLVM_MAJOR_VERSION_21_COMPATIBLE on as…
toudefu Jan 16, 2026
6eb11bc
[BUILD] Update the git path for ascendnpu-ir (#297)
CHAIYD-2025 Jan 21, 2026
0078758
[CI] Fix check docs only on branch triton_v3.2.x (#308)
zhzhcookie Jan 23, 2026
9d86ee6
[BUILD] Add FlagtreeConfig class (#302)
SabYic Jan 27, 2026
b34a886
[SPEC] Fix ascend spec (#324)
zhzhcookie Jan 28, 2026
db585cd
[BACKEND] Update ascend backend to 29d243e 20260129 (#335)
zhzhcookie Feb 11, 2026
663e0d1
[SPEC] Update ascend import path (#358)
zhzhcookie Feb 13, 2026
4f95b31
[VERSION] Update version to 0.4.1 on triton_v3.2.x branch (#359)
zhzhcookie Feb 15, 2026
c0b7254
[BUILD] Add unpack .triton shell for offline-building on triton_v3.2.…
zhzhcookie Feb 18, 2026
232ae95
[Build] fix editable install bugs on ascend backend (#375)
Galaxy1458 Feb 27, 2026
dd0f736
[CI]Add editable mode install ci (#378)
Galaxy1458 Feb 28, 2026
42e650a
[CI] Triton v3.2.x fix action (#391)
git-flyer Mar 10, 2026
837fe3e
[TLE] Add triton language extension (tle) support on ascend backend (…
small-cat Mar 16, 2026
366c432
[VERSION] Update flagtree version on triton_v3.2.x (#451)
zhzhcookie Mar 19, 2026
b3091e9
[CI] Triton_v3.2.x Changed all the yml files in the workflows folder …
git-flyer Mar 19, 2026
4d9e18e
[TLE] Add prefill and decode triton kernels for causal conv1d (#462)
small-cat Mar 20, 2026
e4660c5
[Docs] Add Ascend NPU installation guide for FlagTree (#484)
lindylin1817 Mar 30, 2026
d2f275b
[CI] triton_v3.2.x modify the code-format-check.yml file to use the a…
git-flyer Mar 31, 2026
7fcddf4
[CI] Update build and test files on triton_v3.2.x branch to use check…
git-flyer Apr 12, 2026
de9f0cf
[Build] Update package homepage toflagtree on triton_v3.2.x (#517)
Stardep-lmc Apr 15, 2026
f5366fb
[CI] Triton_v3.2.x fix ascend and nvidia backend yml file bugs (#524)
git-flyer Apr 17, 2026
1d7126d
[VERSION] Update flagtree whl version on triton_v3.2.x (#533)
zhzhcookie Apr 21, 2026
75d6117
[BACKEND] Initializing the mthreads backend in version triton3.2.x (#…
Kylin1207 Apr 23, 2026
eb04eb7
[BUILD] [CI/CD] Add mthreads sdk and cicd (#544)
zhzhcookie Apr 23, 2026
2d9e06c
[BACKEND] Update mtcc on mthreads3.2 (#549)
zhzhcookie Apr 24, 2026
a651019
[CI] triton_v3.2.x, add flagcicd-nv-build-and-test.yml (#554)
git-flyer Apr 29, 2026
35cc929
[TLE] Enable cvpipeline for sfa (#561)
huanghaoXcore Apr 29, 2026
bd8f032
[TLE] Add cv pipeline design doc and example (#590)
huanghaoXcore May 18, 2026
b238fdb
[TLE] Optimize causal_conv1d_prefill tiling for small dim to reduce o…
Quamaly May 20, 2026
db7e7ff
[TLE] Expose Ascend TLE pipeline support (#599)
huanghaoXcore May 20, 2026
fd29968
[Ascend] Align do_bench_npu with CUDA's do_bench (#582)
wangziy1 May 20, 2026
5e56a25
[Ascend] [Autotuner] Fix ascend autotuner to do_bench_npu (#605)
zhzhcookie May 21, 2026
4c30700
[TLE] [DOC] Add TLE interface documentation (#608)
yanminghui123 May 22, 2026
f6f2d58
[TLE] [DOC] Add TLE interface documentation (#614)
yanminghui123 May 26, 2026
28c75a9
[TLE] Fix sync issues in do_bench_npu (#634)
wangziy1 Jun 3, 2026
b522050
[TLE] Add tle_swiglu kernel on Ascend NPU (#631)
wangziy1 Jun 3, 2026
cb91251
[TLE] [DOC] Add TLE DSA interface documentation (#620)
yanminghui123 Jun 3, 2026
4c1f9a9
[TLE-Struct] Add fused RMSNorm and RoPE kernel (#598)
yanminghui123 Jun 4, 2026
fefa349
[TLE] Add AddRmsNormBias triton kernel on ascend (#641)
small-cat Jun 7, 2026
a53e16e
[VERSION] Upgrade to flagtree v0.6.0 (#661)
zhzhcookie Jun 7, 2026
62993ac
[Ascend][CI/CD] Change label and fix default platform (#662)
zhzhcookie Jun 7, 2026
951af63
[TLE][DOC] Add TLE compile_hint and interface documentatioin. (#675)
wangziy1 Jun 11, 2026
f307db7
Initialize Common IR
kateyijian Jun 16, 2026
ca2e8c5
Initialize Common IR
kateyijian Jun 16, 2026
d033d64
Add DSL extension
kateyijian Jun 18, 2026
6b8c19d
Add build script
kateyijian Jun 22, 2026
029660e
fix compiler error
kateyijian Jun 23, 2026
deb05c4
Add compiler README file
kateyijian Jun 23, 2026
c98b356
Add the conversion tileir to hivm ir
kateyijian Jun 24, 2026
5b18715
CommonIR: emit TileIR buffer ops from TLE DSA frontend
macqueen09 Jun 25, 2026
d7eee6a
CommonIR: add buffer subview frontend support
macqueen09 Jun 27, 2026
2f77617
tileir: convert copy(GM->X) to memref.copy
Jun 29, 2026
4b197f3
Rewrite FA Op
kateyijian Jun 30, 2026
7296c4b
Fa op fix (#8)
103yiran Jul 1, 2026
e89a95b
add Fa dummy kernel (#9)
103yiran Jul 1, 2026
67e1d55
Remove redundant copy (#10)
zhongsanming Jul 2, 2026
509d608
Add replace mlir
kateyijian Jul 2, 2026
41e307d
fix(test): fix pv_part name
103yiran Jul 2, 2026
e36d1a4
feat(test): use with
103yiran Jul 2, 2026
1d2e666
fix(test): fix attn_score name
103yiran Jul 2, 2026
2307112
fix(test): remove l0c alloc
103yiran Jul 2, 2026
45d7bf4
fix(third_party): disable ubuf_saving
103yiran Jul 2, 2026
5282950
CommonIR: add scope marker and cleanup TileIR lowering
macqueen09 Jul 1, 2026
3efc53c
feat(test): use scope
103yiran Jul 2, 2026
b4a0eaf
Fix TileIR copy memory-space lowering
macqueen09 Jul 3, 2026
27c9d26
Emit scope op for TLE scope
macqueen09 Jul 3, 2026
6bc90c5
Fix compile and runtime issue
kateyijian Jul 3, 2026
0be1601
Next step to fixx error
kateyijian Jul 3, 2026
879ca44
Common ir poc fix enum import (#14)
zhongsanming Jul 4, 2026
a1887d7
add matmul op (#13)
103yiran Jul 4, 2026
0cf9118
Enable compile flow and Success to run matmul
kateyijian Jul 5, 2026
4368dbb
Matmul add (#15)
103yiran Jul 6, 2026
9c9d632
fix native matmul add residual
Jul 7, 2026
128cc30
Merge pull request #18 from zhongsanming/common-ir-poc-fix-matmul-res…
zhongsanming Jul 7, 2026
09b695d
Native fa (#17)
103yiran Jul 8, 2026
1d1ac39
fix subview lowering issue, now it runs but produces wrong result (#19)
zhongsanming Jul 8, 2026
35726fa
Add compile option
kateyijian Jul 9, 2026
09322f1
make number of slots a constexpr parameter in prefetch (#21)
zhongsanming Jul 10, 2026
74cd66f
add envvar to skip CSE pass (#23)
zhongsanming Jul 10, 2026
245dc28
Add DSA native matmul dump support (#20)
macqueen09 Jul 10, 2026
f571cf7
change number of slot to 3
Jul 13, 2026
34505ef
Add tle.gpu TileIR to TTGIR POC
macqueen09 Jul 16, 2026
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
The table of contents is too big for display.
Diff view
Diff view
  •  
  •  
  •  
41 changes: 4 additions & 37 deletions .github/CODEOWNERS
Original file line number Diff line number Diff line change
Expand Up @@ -2,47 +2,14 @@
# the repo. Unless a later match takes precedence,
# @global-owner1 and @global-owner2 will be requested for
# review when someone opens a pull request.
* @ptillet
* @zhzhcookie

# --------
# Analyses
# Setup
# --------
# Alias analysis
include/triton/Analysis/Alias.h @Jokeren
lib/Analysis/Alias.cpp @Jokeren
# Allocation analysis
include/triton/Analysis/Allocation.h @Jokeren
lib/Analysis/Allocation.cpp @Jokeren
# Membar analysis
include/triton/Analysis/Membar.h @Jokeren
lib/Analysis/Membar.cpp @Jokeren
# AxisInfo analysis
include/triton/Analysis/AxisInfo.h @ptillet
lib/Analysis/AxisInfo.cpp @ptillet
# Utilities
include/triton/Analysis/Utility.h @Jokeren
lib/Analysis/Utility.cpp @Jokeren

# ----------
# Dialects
# ----------
# Pipeline pass
lib/Dialect/TritonGPU/Transforms/Pipeline.cpp @ptillet
# Prefetch pass
lib/Dialect/TritonGPU/Transforms/Prefetch.cpp @ptillet
# Coalesce pass
lib/Dialect/TritonGPU/Transforms/Coalesce.cpp @ptillet
# Layout simplification pass
lib/Dialect/TritonGPU/Transforms/Combine.cpp @ptillet

# -----------
# Conversions
# -----------
# TritonToTritonGPU
include/triton/Conversion/TritonToTritonGPU/ @ptillet
lib/Dialect/TritonGPU/Transforms/TritonGPUConversion.cpp @ptillet
python/setup.py @Galaxy1458
python/setup_tools/** @Galaxy1458

# -----------
# third_party
# -----------
third_party/amd/ @antiagainst @zhanglx13
43 changes: 16 additions & 27 deletions .github/PULL_REQUEST_TEMPLATE.md
Original file line number Diff line number Diff line change
@@ -1,27 +1,16 @@
The core Triton is a small number of people, and we receive many PRs (thank
you!). To help us review your code more quickly, **if you are a new
contributor (less than 3 PRs merged) we ask that you complete the following
tasks and include the filled-out checklist in your PR description.**

Complete the following tasks before sending your PR, and replace `[ ]` with
`[x]` to indicate you have done them.

- [ ] I am not making a trivial change, such as fixing a typo in a comment.

- [ ] I have written a PR description following these
[rules](https://cbea.ms/git-commit/#why-not-how).

- [ ] I have run `pre-commit run --from-ref origin/main --to-ref HEAD`.

- Select one of the following.
- [ ] I have added tests.
- `/test` for `lit` tests
- `/unittest` for C++ tests
- `/python/test` for end-to-end tests
- [ ] This PR does not need a test because `FILL THIS IN`.

- Select one of the following.
- [ ] I have not added any `lit` tests.
- [ ] The `lit` tests I have added follow these [best practices](https://mlir.llvm.org/getting_started/TestingGuide/#filecheck-best-practices),
including the "tests should be minimal" section. (Usually running Python code
and using the instructions it generates is not minimal.)
<!-- PR Title
[component] brief description
component options:
- BUILD
- CI/CD
- DOC
- FRONTEND
- BACKEND
- AUTOTUNER
- CACHE
- LAYOUTS
- PIPELINE
- PROTON
- TEST
- OTHER
-->
105 changes: 105 additions & 0 deletions .github/workflows/ascend-build-and-test.yml
Original file line number Diff line number Diff line change
@@ -0,0 +1,105 @@
name: Ascend3.2-Build-And-Test

on:
push:
branches: [ "triton_v3.2.x" ]
pull_request:
branches: [ "triton_v3.2.x" ]

concurrency:
group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref }}
cancel-in-progress: true

jobs:
ascend-build-and-test:
runs-on: ascend3.2
if: ${{ github.repository == 'FlagTree/flagtree' || github.repository == 'flagos-ai/flagtree' }}
steps:
- name: Setup environment
shell: bash
run: |
source ~/env.sh
env | grep -E '^(http_proxy|https_proxy|all_proxy|no_proxy)=' >> $GITHUB_ENV || true

- name: Smart Checkout
uses: flagos-ai/FlagTree/.github/actions/smart-checkout@main
with:
checkout_version: 'v6'

- name: Check if backend-relevant files changed
id: check_backend
uses: flagos-ai/FlagTree/.github/actions/check-backend-changed@main
with:
backend: ascend

- name: FlagTree Editable Build And Test on Ascend
if: steps.check_backend.outputs.should_skip != 'true'
shell: bash
run: |
set -x
export FLAGTREE_BACKEND=ascend
source ~/env.sh
python3 -m pip uninstall flagtree -y
pushd python
MAX_JOBS=32 python3 -m pip install -e . --no-build-isolation
popd
source /usr/local/Ascend/ascend-toolkit/set_env.sh
python3 -c "import triton"
python3 -m pip uninstall flagtree -y

- name: FlagTree Build on Ascend
if: steps.check_backend.outputs.should_skip != 'true'
shell: bash
run: |
set -x
export FLAGTREE_BACKEND=ascend
source ~/env.sh
cd python
MAX_JOBS=32 python3 -m pip install . --no-build-isolation

- name: FlagTree Test on Ascend
if: steps.check_backend.outputs.should_skip != 'true'
shell: bash
run: |
set -x
source /usr/local/Ascend/ascend-toolkit/set_env.sh
# tutorials
pushd third_party/ascend/tutorials
python3 01-vector-add.py
python3 02-fused-softmax.py
python3 03-layer-norm.py
python3 04-fused-attention.py
python3 05-matrix-multiplication.py
python3 06-demo-autotune.py
python3 07-profiler.py
python3 08-demo-libentry.py
python3 09-gather.py >output-09-gather.log
python3 10-gather_sorted.py >output-10-gather_sorted.log
python3 11-rab_time.py
#python3 12-hstu_attention.py
python3 13-matrix-multiplication-optimized.py
python3 14-accuracy-comparison.py
#python3 15-embedding_gather_demo.py
popd
# pytest_ut
pushd third_party/ascend/unittest/pytest_ut
python3 -m pytest . \
--ignore=test_add_mindspore.py \
--ignore=test_softmax_mindspore.py \
--ignore=test_assume.py \
--ignore=test_index_select.py
popd
# flagtree tle test
pushd python/test/tle
python3 test_vec_add.py
python3 test_vec_add_2d.py
python3 test_vec_add_mix.py
python3 test_vec_mathOps.py
python3 test_tle_with_hints.py
popd
# flagtree tle tutorial
pushd python/tutorials/tle
python3 06-swiglu.py
python3 07-kv-rmsnorm-rope-cache.py
pytest -s -k "accuracy" 08-add-rms-norm-bias.py
popd
44 changes: 44 additions & 0 deletions .github/workflows/ascend3.2-delivery.yml
Original file line number Diff line number Diff line change
@@ -0,0 +1,44 @@
name: CD-Ascend3.2

on:
workflow_dispatch:
inputs:
WHL_VER:
description: '0.6.0.dev1 < 0.6.0a1 < 0.6.0b1 < 0.6.0rc1 < 0.6.0 < 0.6.0.post1'
required: true
default: '0.6.0'
PYTHON_VER:
required: true
default: '3.11'
PLATFORM:
required: true
default: 'aarch64'
FLAGTREE_PYPI_KEY:
required: false
workflow_call:
inputs:
WHL_VER:
type: string
required: true
default: '0.6.0'
PYTHON_VER:
type: string
required: true
default: '3.11'
PLATFORM:
type: string
required: true
default: 'aarch64'
FLAGTREE_PYPI_KEY:
type: string
required: false

jobs:
call-delivery:
uses: flagos-ai/flagtree/.github/workflows/ascend3.2-delivery.yml@main
with:
WHL_VER: ${{ inputs.WHL_VER }}
PYTHON_VER: ${{ inputs.PYTHON_VER }}
PLATFORM: ${{ inputs.PLATFORM }}
FLAGTREE_PYPI_KEY: ${{ inputs.FLAGTREE_PYPI_KEY }}
secrets: inherit
36 changes: 36 additions & 0 deletions .github/workflows/cambricon-build-and-test.yml
Original file line number Diff line number Diff line change
@@ -0,0 +1,36 @@
name: Cambricon-Build-And-Test

on:
push:
branches: [ "main" ]
pull_request:
branches: [ "main" ]

concurrency:
group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref }}
cancel-in-progress: true

jobs:
cambricon-build-and-test:
runs-on: cambricon
if: ${{ github.repository == 'FlagTree/flagtree' || github.repository == 'flagos-ai/flagtree' }}
steps:
- name: Setup environment
shell: bash
run: |
source ~/env.sh
env | grep -E '^(http_proxy|https_proxy|all_proxy|no_proxy)=' >> $GITHUB_ENV || true

- name: Smart Checkout
uses: flagos-ai/FlagTree/.github/actions/smart-checkout@main
with:
checkout_version: 'v6'

- name: FlagTree Build on Cambricon
shell: bash
run: |
set -x
pip uninstall -y triton
export FLAGTREE_BACKEND=cambricon
cd python
MAX_JOBS=8 python3 -m pip install . --no-build-isolation
25 changes: 25 additions & 0 deletions .github/workflows/code-format-check.yml
Original file line number Diff line number Diff line change
@@ -0,0 +1,25 @@
name: Code-Format-Check

on:
schedule:
- cron: '0 21 * * *'
push:
branches: [ "main", "triton_v3.2.x", "triton_v3.3.x", "triton_v3.4.x", "triton_v3.5.x"]
pull_request:
branches: [ "main", "triton_v3.2.x", "triton_v3.3.x", "triton_v3.4.x", "triton_v3.5.x"]

concurrency:
group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref }}
cancel-in-progress: true

jobs:
pre-commit:
runs-on: ubuntu-latest
steps:
- name: Checkout code
id: checkout
uses: actions/checkout@v6
with:
token: ${{ secrets.BOT_PAT_TOKEN || github.token }}
- name: Code format check
uses: flagos-ai/FlagTree/.github/actions/code-format-check@main
57 changes: 57 additions & 0 deletions .github/workflows/code_scan.yml
Original file line number Diff line number Diff line change
@@ -0,0 +1,57 @@
name: code-scan

on:
push:
branches: [ "main" ]
pull_request:
branches: [ "main" ]

jobs:
scan-code-and-report:
runs-on: scan
if: ${{ github.repository == 'FlagTree/flagtree' }}
concurrency:
group: scan-code-and-report-${{ github.event.pull_request.number || github.ref }}
cancel-in-progress: true

steps:
- name: Setup environment
shell: bash
run: |
source ~/env.sh
env | grep -E '^(http_proxy|https_proxy|all_proxy|no_proxy)=' >> $GITHUB_ENV || true

- name: Smart Checkout
uses: flagos-ai/FlagTree/.github/actions/smart-checkout@main
with:
checkout_version: 'v6'

- name: Scan flagtree repo
shell: bash
run: |
set -x
git config --global --add safe.directory ../flagtree
if [ "${{ github.event_name }}" == "pull_request" ]; then
echo "This is a pull request event. PR number is ${{ github.event.pull_request.number }}"
PR_ID=${{ github.event.pull_request.number }}
elif [ "${{ github.event_name }}" == "push" ]; then
PR_NUMBER=$(git log -1 --pretty=format:'%s' | grep -oE '#[0-9]+' | grep -oE '[0-9]+')
echo "This is a push event. The relate PR number is ${PR_NUMBER}"
PR_ID=${PR_NUMBER}
fi

python /work/flag_tree/code_can/scan_code.py --pr ${PR_ID} --hash ${GITHUB_SHA} --attempt ${GITHUB_RUN_ATTEMPT}

- name: Code Scan Report
shell: bash
run: |
set -x
git config --global --add safe.directory ../flagtree
if [ "${{ github.event_name }}" == "pull_request" ]; then
PR_ID=${{ github.event.pull_request.number }}
elif [ "${{ github.event_name }}" == "push" ]; then
PR_NUMBER=$(git log -1 --pretty=format:'%s' | grep -oE '#[0-9]+' | grep -oE '[0-9]+')
PR_ID=${PR_NUMBER}
fi
echo "Repo Scan report:"
echo "http://120.92.44.177/tree_scan_report/PR${PR_ID}-${GITHUB_SHA}-${GITHUB_RUN_ATTEMPT}.html"
Loading