-
Notifications
You must be signed in to change notification settings - Fork 36
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.
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
Author
Label
Projects
Milestones
Assignee
Sort
Issues list
[Bug] Code bug in transpose 2D block IO lowering
bug
Something isn't working
codegen: mlir
performance
Reduce GenISA usages
codegen: mlir
dependencies: igc
driver: rolling
enhancement
New feature or request
Deprecate fast prefetch implementation
codegen: mlir
dependencies: igc
driver: rolling
enhancement
New feature or request
performance
What's version of Intel triton? or "No module named 'triton.ops'"
question
Further information is requested
tests: ecosystem
Align resulting computation of GBs, TFLOPs in tutorials
enhancement
New feature or request
tests: tutorials
upstream: triton
Incorrect ylabel in tutorial 06-fused-attention
enhancement
New feature or request
tests: tutorials
upstream: triton
Port "sub-group transpose reduction" to default path
codegen: attention
enhancement
New feature or request
performance
Investigate failures from updating SPIRV-LLVM-Translator
bug
Something isn't working
dependencies: spirv_translator
tests: ut
Subset of E2E models for accuracy tests
accuracy
ci
enhancement
New feature or request
tests: e2e
validation
Merge OpenAI Triton till Sept 27th
enhancement
New feature or request
upstream: rebase
PR to be up-streamed
[sycl-free-inference-for-llms] integrate triton gemm/attention in pytorch for xpu
enhancement
New feature or request
[sycl-free-inference-for-llms] flashattention kernel get up to ~100tflops
enhancement
New feature or request
performance
[sycl-free-inference-for-llms] gemm kernel perf get up to ~290tflops
enhancement
New feature or request
performance
Previous Next
ProTip!
Type g i on any issue or pull request to go back to the issue listing page.