Enhance XPU support for benchmarks, profiling, and verification - #11

Open

sgurwinderr wants to merge 7 commits into

RightNow-AI:mainfrom

sgurwinderr:main

Commits on Apr 14, 2026

Add comprehensive XPU support for benchmark/profile/verify flows
gurwinderintel
committed
c2d3044
Update prepare.py
Gurwinder Singh
authored
5ffcb5a
Polish XPU upstream patch and docs
gurwinderintel
committed
710636c

Commits on Apr 22, 2026

exp 3: num_warps=8 for reduce kernel on XPU for more parallelism
gurwinderintel
committed
f33edfd
fused_mlp baseline: relax bfloat16 tolerance to 0.1 for XPU numerical range
gurwinderintel
committed
b6aa861

Commits on May 4, 2026

Unified OOM exception handling for CUDA/XPU
gurwinderintel
committed
51e3147
Revert README to pre-XPU docs version
gurwinderintel
committed
dc82327