]> git.djapps.eu Git - pkg/ggml/sources/ggml/commit
overview / pkg / ggml / sources / ggml / commit
summary | shortlog | log | commit | commitdiff | tree
(parent: 86a110c)
cuda : revert CUDA_SCALE_LAUNCH_QUEUES override until investigated (llama/19227)
authorGaurav Garg <redacted>
Tue, 3 Feb 2026 06:41:02 +0000 (12:11 +0530)
committerGeorgi Gerganov <redacted>
Sat, 7 Feb 2026 08:37:38 +0000 (10:37 +0200)
commit8c2086e7fe98fd222835c12f41358a1f7715f6d3
treeb4f45030eb57717030f8a08cc44c014d00ea6e34tree
parent86a110caf3327e07ec3950119c64fb6a53159ceccommit | diff
cuda : revert CUDA_SCALE_LAUNCH_QUEUES override until investigated (llama/19227)

Hangs were reported on Jetson Orin AGX if we set CUDA_SCALE_LAUNCH_QUEUES=4x. Reverting the previous PR (#19042) and updating the document to consider setting CUDA_SCALE_LAUNCH_QUEUES=4x for faster throughput on multi-GPU systems.
src/ggml-cuda/ggml-cuda.cu diff | blob | history
Packaging of ggml-org/ggml