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

180 Open 1,241 Closed
180 Open 1,241 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

[SGLANG] Initial enabling and its integration into the test scripts local builds tests: ecosystem
#3623 opened Mar 5, 2025 by vlad-penkin 5. Ecosystem enabling and performance
[SGLANG][Triton 3.4] Functional enabling and performance benchmarking tests: ecosystem
#3622 opened Mar 5, 2025 by vlad-penkin 5. Ecosystem enabling and performance
1
[Liger-Kernel] Initial enabling and its integration into the test scripts local builds tests: ecosystem
#3621 opened Mar 5, 2025 by vlad-penkin 5. Ecosystem enabling and performance
[FlagGems] Initial enabling and its integration into the test scripts local builds tests: ecosystem
#3620 opened Mar 5, 2025 by vlad-penkin 5. Ecosystem enabling and performance
@LiyangLingIntel
Enhance FlashAttention and FlexAttention benchmarks to evaluate Decode and Append stages
#3617 opened Mar 5, 2025 by mfrancepillois
Enhance FlashAttention and FlexAttention benchmarks with paged Attention codegen: attention performance
#3616 opened Mar 5, 2025 by mfrancepillois 4. [Performance] Core
Enhance FlashAttention and FlexAttention benchmarks with GQA codegen: attention performance
#3615 opened Mar 5, 2025 by mfrancepillois 4. [Performance] Core
1
[SGLANG] group quant UT failed tests: ecosystem
#3613 opened Mar 5, 2025 by airMeng 5. Ecosystem enabling and performance
1
Support get global scratch space on GLM code quality codegen: mlir upstream: triton
#3612 opened Mar 5, 2025 by ESI-SYD 2. [Triton] Upstream
Adapt tools/test_aot.py for Windows tests: torchinductor windows
#3611 opened Mar 4, 2025 by anmyachev 1. [PT 2.7 Upstream] TorchInductor
@anmyachev
[Triton 3.5 dev cycle] Fix UT and Tutorials failures tests: tutorials tests: ut
#3608 opened Mar 4, 2025 by vlad-penkin 3. [Triton] Language and Runtime
Fix int8 arith::CmpI with 0xFF issue for DecomposeScaledBlocked pass codegen: mlir tests: ut
#3605 opened Mar 4, 2025 by LiyangLingIntel 3. [Triton] Language and Runtime
@LiyangLingIntel
[PT][CI] Add additional Inductor tests to the defaults for the Inductor UT WF ci dependencies: pytorch tests: torchinductor upstream: pytorch
#3597 opened Mar 3, 2025 by vlad-penkin 1. [PT 2.7 Upstream] TorchInductor
@anmyachev
1
[Loop Specialization]: Specialize loops containing masked operations with loop invariant mask codegen: mlir performance
#3594 opened Mar 3, 2025 by etiotto 4. [Performance] Core
1
@etiotto
[GEMM and FA] Eliminate remaining differences between default and advanced path codegen: attention codegen: gemm performance
#3593 opened Mar 3, 2025 by vlad-penkin 4. [Performance] Core
[CI] Performance benchmarking enhancements ci performance
#3592 opened Mar 3, 2025 by vlad-penkin 7. CI and Tooling
[WIN][CI] Enable regular PyTorch E2E CI runs on Windows ci hw: bmg tests: e2e windows
#3591 opened Mar 3, 2025 by vlad-penkin 7. CI and Tooling
@pbchekin
[WIN][CI] Enable regular PyTorch UT CI runs on Windows ci hw: bmg tests: torchinductor windows
#3590 opened Mar 3, 2025 by vlad-penkin 7. CI and Tooling
@pbchekin
[PyTorch Upstream] inductor/test_torchinductor_opinfo.py::TestInductorOpInfoXPU::test_comprehensive_masked_std_xpu_int64 regression tests: torchinductor upstream: pytorch
#3589 opened Mar 3, 2025 by anmyachev 1. [PT 2.7 Upstream] TorchInductor
1
@anmyachev
3
[FlexAttention] FlexDecoding accuracy discrepancy between XPU and CUDA while compiling torch.ops.higher_order.flex_attention accuracy upstream: pytorch
#3588 opened Mar 3, 2025 by hoshibara 1. [PT 2.7 Upstream] TorchInductor
[HF] and [Liger-kernels][FLCE] giving error since 02/27/2025 community tests: e2e upstream: pytorch
#3585 opened Mar 1, 2025 by mgrabban 3. [Triton] Language and Runtime
@alexbaden
1
[PT 2.8] Enable Inductor UT and E2E dependencies: pytorch upstream: pytorch
#3583 opened Feb 28, 2025 by vlad-penkin 3. [Triton] Language and Runtime
Enable Triton FA bwd kernel benchmark on BMG hw: bmg performance
#3580 opened Feb 28, 2025 by anmyachev 3. [Triton] Language and Runtime
Investigate how to fix ElementwiseOpConversionBase::maybeDeduplicate failing for DpasEncoding codegen: mlir research upstream: rebase PR to be up-streamed
#3577 opened Feb 28, 2025 by dchigarev 3. [Triton] Language and Runtime
Cat+BN+Relu inference is very slow on Max 1550 performance upstream: pytorch
#3576 opened Feb 28, 2025 by jianyizh 4. [Performance] Core
2
ProTip! Updated in the last three days: updated:>2025-03-02.