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
Tutorial 09-experimental-block-pointer with enabled block pointer fails with LTS driver (803.45)
bug
Something isn't working
#1159
opened May 20, 2024 by
pbchekin
Rewrite RewriteTensorPointer pass to support 2D block store
codegen: dpas
enhancement
New feature or request
performance
[Upstream] Upstream the threads_per_warp changes in #414
enhancement
New feature or request
upstream: triton
Should we have a microbenchmark for FP conversion operations
enhancement
New feature or request
performance
question
Further information is requested
Tutorial example 03 performance issue
bug
Something isn't working
community
performance
tests: tutorials
Lowering Something isn't working
codegen: spirv
arith::TruncFOp
from fp32 to bf16 is harder than it should be.
accuracy
bug
Unskipping test_flash_attention.py breaks regular runs
bug
Something isn't working
ci
tests: tutorials
Execution hang for float16 type with MHA block
bug
Something isn't working
codegen: attention
tests: e2e
[#6 GEMM Performance] gemm perf of big K shape reach 80% of XeTLA
codegen: gemm
enhancement
New feature or request
performance
Previous Next
ProTip!
Adding no:label will show everything without a label.