#nixl

The latest AI and developer news about #nixl, with the original source and a short summary.

vLLM Introduces Push-Based KV Transfer to Cut TTFT by Up to 3x in Disaggregated Inference

The new NIXL connector-based push method reduces Time To First Token by up to 3x and Time Per Output Token by up to 30% compared to pull-based approaches.

·2·2026.10.14 09:00

ModelExpress: Deploying Model Artifacts at the Speed of Light

NVIDIA ModelExpress dramatically reduces model weight loading times through direct GPU-to-GPU data transfer.

·2026.07.27 09:00

The Heap Lies: Debugging a vLLM Memory Leak

Mistral AI shared the technical process of tracking down a memory leak that occurred in a distributed serving environment of vLLM.

·1·2026.01.21 09:00