Repository navigation

glm cuda: TF_GLM_CHUNK_SOLO takes two grid chunks in one pass while no other stream decodes - #1823

Merged

ashhart merged 3 commits into

ashhart:mainfrom

Bizuayeu:pr/glm-chunk-solo

Nov 1, 2026

Merged

glm cuda: TF_GLM_CHUNK_SOLO takes two grid chunks in one pass while no other stream decodes#1823
ashhart merged 3 commits into
ashhart:mainfrom
Bizuayeu:pr/glm-chunk-solo

Commits on Nov 1, 2026
- glm cuda: TF_GLM_CHUNK_SOLO takes two grid chunks in one pass while no other stream decodes
  
  Bizuayeu
  authored and
  ashhart
  committed
  acfc648
- docs: the GLM CUDA recipe lists TF_GLM_CHUNK_SOLO with its measured trade-off
  
  Bizuayeu
  authored and
  ashhart
  committed
  00b3cbb
- changelog: #1823 under Unreleased
  ashhart
  committed
  858f2df