Skip to content

Issues: intel/intel-xpu-backend-for-triton

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

135 Open 839 Closed
135 Open 839 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Assignee
Filter by who’s assigned
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Issues list

The workflow "Automate update Translator cid" cannot create PRs bug Something isn't working ci
#2427 opened Oct 5, 2024 by pbchekin 7. CI
@AshburnLee
[Benchmarks] Add microbenchmark with A^t@B enhancement New feature or request performance
#2424 opened Oct 4, 2024 by Egor-Krivov 4.3 [Performance] Tracking
@Egor-Krivov
Add dotOperandDpasToLinearLayout (Tensor B) codegen: dpas enhancement New feature or request
#2423 opened Oct 3, 2024 by whitneywhtsang 0.3 [Triton] Language and Runtime
1
@whitneywhtsang
[Triton Upstream] AttrsDescriptor class refactoring breaks Torch Inductor bug Something isn't working tests: e2e
#2421 opened Oct 3, 2024 by whitneywhtsang 1.0 [UT and Tutorials][Triton 3.0] Pass rate
@alexbaden
[Benchmarks] XeTLA kernel python library build enhancement New feature or request performance
#2418 opened Oct 3, 2024 by vlad-penkin 4.3 [Performance] Tracking
1
@ZzEeKkAa
[PoC] Windows native Triton XPU build enhancement New feature or request research
#2407 opened Oct 2, 2024 by vlad-penkin 5.0 [DGPU] Initial enabling
@gshimansky
[Pytorch pin update] e9a55b4 - inductor/test_triton_kernels.py::CustomOpTests::test_autotune_unbacked test failure bug Something isn't working nightly failure failure in nightly actions tests: torchinductor
#2404 opened Oct 2, 2024 by vlad-penkin 0.1 [PT 2.6 Upstream] TorchInductor
@anmyachev
1
Replace subgroup query operations in TritonGEN with their GPU dialect equivalent codegen: mlir enhancement New feature or request upstream: triton
#2403 opened Oct 2, 2024 by FMarno 0.3 [Triton] Language and Runtime
@FMarno
Investigate commonality between AOT Inductor/Triton/SPIRV Runner and refactor SYCL kernel launcher codegen: spirv enhancement New feature or request research
#2395 opened Oct 1, 2024 by kballeda 0.3 [Triton] Language and Runtime
@alexbaden
1
[XeTLA] Add StreamK and SplitK implementation codegen: gemm enhancement New feature or request performance
#2390 opened Oct 1, 2024 by whitneywhtsang 4.3 [Performance] Tracking
@ESI-SYD
Merge OpenAI Triton till Oct 11st enhancement New feature or request upstream: rebase PR to be up-streamed
#2383 opened Sep 30, 2024 by whitneywhtsang 0.2 [Triton] Upstream
@whitneywhtsang
1
Clean up duplication between MaterializeBlockPtr and RewriteTensorPointer codegen: mlir enhancement New feature or request
#2380 opened Sep 27, 2024 by alexbaden 0.3 [Triton] Language and Runtime
1
@alexbaden
Improve out-of-box performance for GEMM kernels variants codegen: gemm enhancement New feature or request performance umbrella
#2379 opened Sep 27, 2024 by etiotto 4.0 [Performance] Core
@etiotto
2
Assertion error on gemm_splitk_benchmark.py bug Something isn't working tests: ut
#2377 opened Sep 27, 2024 by etiotto 1.0 [UT and Tutorials][Triton 3.0] Pass rate
@LiyangLingIntel
2
[Research]: Evaluate enhanced AxisAnalysis for block ptr materialization codegen: mlir research
#2375 opened Sep 27, 2024 by etiotto 0.3 [Triton] Language and Runtime
@sommerlukas
1
Improve performance of tt.load and tt.store for FP8 when converting block ptr to regular ptrs codegen: attention codegen: gemm performance
#2374 opened Sep 27, 2024 by etiotto 0.3 [Triton] Language and Runtime
@etiotto
[Performance] Investigate and optimize GEMM StreamK performance codegen: gemm performance
#2368 opened Sep 27, 2024 by LiyangLingIntel 4.0 [Performance] Core
@LiyangLingIntel
02-fused-softmax: PyTorch faster than Triton on Max 1550 in composite mode performance research
#2363 opened Sep 26, 2024 by pbchekin 4.3 [Performance] Tracking
2
02-fused-softmax: differences for Max 1100 and Max 1550 performance research tests: tutorials
#2362 opened Sep 26, 2024 by pbchekin 4.3 [Performance] Tracking
@alexbaden
[Performance] The GEMM performance with the column major B matrix is not as good as row major B matrix. enhancement New feature or request performance
#2354 opened Sep 26, 2024 by chengjunlu 4.0 [Performance] Core
1
Allocation analysis only consider NvidiaMmaEncodingAttr codegen: mlir enhancement New feature or request
#2340 opened Sep 25, 2024 by LiyangLingIntel 0.3 [Triton] Language and Runtime
[FA2 performance] flashattention with dim=128 get ~90% of xetla codegen: attention enhancement New feature or request performance
#2339 opened Sep 25, 2024 by Dewei-Wang-sh 4.0 [Performance] Core
@Dewei-Wang-sh
1
Add uninstall note to README documentation Improvements or additions to documentation
#2326 opened Sep 24, 2024 by alexbaden 9. DX/UX
[Performance] Check the possibility that improve Triton atomic op performance. performance research
#2322 opened Sep 24, 2024 by LiyangLingIntel 4.0 [Performance] Core
@LiyangLingIntel
[Performance] Register spill in flash attention forward kernel codegen: attention dependencies: igc enhancement New feature or request performance
#2317 opened Sep 24, 2024 by chengjunlu 4.0 [Performance] Core
1
@chengjunlu
1
ProTip! no:milestone will show everything without a milestone.