Record · corpus.blog/posts/01a087db-5e81-7342-a556-22dd1597754a

Auto-research with codex: How I achieved a 232x Faster Kernel over baseline with Codex in GPU Mode's qr_v2 problem

sankalp's blog · published 8 July 2026

A record is what was published and who pointed at it. The text of the post is not held here: read it at the source, or at its Wayback capture.

Read the post at sankalp.bearblog.dev Wayback capture

0blogs citing

First cited: —
Most recent: —
Rank this month: not ranked
Links from this post: 25
Words captured: 5,592

Cited by

No blog we hold has cited this post yet.

Links from this post

Link

Host

https://www.gpumode.com/leaderboard/774?tab=rankings

gpumode.com

https://www.gpumode.com/news/linear-algebra-kernels-age-of-research

gpumode.com

https://ml-mike.com/writing/qr_v2/

ml-mike.com

https://github.com/karpathy/autoresearch/blob/master/program.md

github.com

https://t.co/imqZc6RPkz

t.co

https://x.com/dejavucoder/status/2070965828829564946?ref_src=twsrc%5Etfw

x.com

https://pytorch.org/blog/why-is-pytorch-compile-so-fast-kernel-fusion/

pytorch.org

one must imagine sisyphus full of ideas: on exercising the idea muscle Read at the source ↗

sankalp.bearblog.dev

https://t.co/fRkegyMs5t

t.co

https://x.com/claudeai/status/2042308622181339453?ref_src=twsrc%5Etfw

x.com

https://www.gpumode.com/leaderboard/775?tab=rankings

gpumode.com

https://modal.com/gpu-glossary

modal.com

How to Optimize a CUDA Matmul Kernel for cuBLAS-like Performance: a Worklog Read at the source ↗

siboehm.com

https://www.paulwillchan.com/articles/outperforming-cublas-b200

paulwillchan.com

Outperforming cuBLAS on H100: a Worklog Read at the source ↗

cudaforfun.substack.com

https://veitner.bearblog.dev/blog/

veitner.bearblog.dev

https://gau-nernst.github.io/

gau-nernst.github.io

https://x.com/marksaroufim

x.com

https://x.com/_arohan_

x.com

https://x.com/myainotez

x.com

https://x.com/Michael_J_Lutz

x.com

https://x.com/levidiamode

x.com

https://tokenbender.com/

tokenbender.com

https://x.com/blelbach

x.com

https://t.co/Yx6EpH5PKo

t.co