Record · pytorch.org

TLX Block Attention: A Warp-Specialized Blackwell Kernel for Fixed-Block Sparse Self-Attention

PyTorch Blog · 26 May 2026 · 15 min read

A post on pytorch.org, published 26 May 2026, is cited by 1 distinct blog to date (corpus.blog, measured 12 October 2026).

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

Read at pytorch.org Wayback capture

1blog citing

Citing blogs and links by period
Period	Blogs	Links
All time	1	1
Last 90 days	1	1
Last 30 days	0	0
Last 7 days	0	0

Figures for this page
First cited	3 Aug 2026
Latest citation	3 Aug 2026
Words blogs link it with	1
External links	13
Length	3,392

Words blogs link it with

The words blogs link this page with, by how many different blogs use them.

reduced long-sequence self-attention latency by up to 68% with neutral ne (normalized entropy, a model-quality metric)1 blog

Cited by 1 blog in 1 post

One row per blog, the most recently citing first. Each shows the blog's newest citing post; open a row for the others.

Blog

Newest citing post

Date

Engineering at Meta1 post

GEM Training: How Meta Doubled the Efficiency of Its LLM-Scale Ads Foundation Model

linked as “reduced long-sequence self-attention latency by up to 68% with neutral NE (normalized entropy, a mod…”

Read at engineering.fb.com

3 Aug 2026

Links from this post

Link

Host

https://github.com/facebookresearch/ads_model_kernel_library Who cites this page

github.com

https://l.facebook.com/l.php?u=https%3A%2F%2Fgithub.com%2Ftriton-lang%2Ftriton-ext&h=AUAOq8QYbTX6IusNi8rHABdpQH9B8xew6KYsjPTIj7t7ziwHNEIGjtsBFtKp-gtiLGDsDq92cpWTOX1ErSS05AY_GKdORcPejIQxMNN3_VT564XkpaSCnYMv-54v3iXi3URUVIBfceAS0uRmvw Who cites this page

l.facebook.com

https://arxiv.org/abs/1911.02972 Who cites this page

arxiv.org

https://l.facebook.com/l.php?u=https%3A%2F%2Fgithub.com%2Ftriton-lang%2Ftriton-ext&h=AUD8yMu0haMw6d272mfnKLyPQGAzLzIYglBbh1LZ4klMCKgd0std74g4lH8r6sst8DVeP6KNWx3T3ZRv0OyeTM7u-Jr2t3Av244-fCsvpiVCJOrDVpPuf3PpZTCJNEkBtTTT-2P7XghPEt25DA Who cites this page

l.facebook.com

https://l.facebook.com/l.php?u=https%3A%2F%2Fgithub.com%2Ffacebookresearch%2Fads_model_kernel_library&h=AUDYkgqWD6TouuZ-i2ffV3BxY3-89i8YVjg3BPQBnGwTMlfo34S4MSMhtmHxAbtqg3BoRQXsxBd4cjypJheyHqrvVEyFfaSC9HV8WIhIe2EpIX0JWnxXqU4dGZsPl52FjJcsItYomzRcZoCWtQ Who cites this page

l.facebook.com

https://github.com/triton-lang/triton-ext Who cites this page

github.com

https://l.facebook.com/l.php?u=https%3A%2F%2Fpytorch.org%2Fblog%2Fflexattention%2F&h=AUAmLdWPy06rJj2mmqzTlhtyLrpJXn3s3lb5f4bumSNv7kkHeIbMyST-3bpdxLl_N-wNnAkFk0dtMcBi3bp0nS6D2LFxQc9U0AoiMIdYMd_aWiCSTG_CNM3jtkJTK1b3fmFofSxcOeUNy7bC8g Who cites this page

l.facebook.com

https://l.facebook.com/l.php?u=https%3A%2F%2Fpytorch.org%2Fblog%2Fwarp-specialization%2F&h=AUAzWimpX4oxkxv-K9ClDKTkrCFzp7J3oyuMEJGCkdW0EAr57OH6DJMFk4t5eIvruWZObmYX8UqewrIkzORaxlnZdsAUsJ2-xejs2_OCYGcLfpEc4rRtpS7K6tmi69b8Bz2Xrx8BeGBIsRWxlw Who cites this page

l.facebook.com

https://arxiv.org/abs/2205.14135 Who cites this page

arxiv.org

https://arxiv.org/abs/2307.08691 Who cites this page

arxiv.org

https://resources.nvidia.com/en-us-blackwell-architecture Who cites this page

resources.nvidia.com

https://www.eecs.harvard.edu/~htk/publication/2019-mapl-tillet-kung-cox.pdf Who cites this page

eecs.harvard.edu

https://arxiv.org/abs/2104.09864 Who cites this page

arxiv.org

TLX Block Attention: A Warp-Specialized Blackwell Kernel for Fixed-Block Sparse Self-Attention

Words blogs link it with

Cited by 1 blog in 1 post

Links from this post

Similar posts