#hip

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

llama.cpp PR #29353 introduces GDN chunked kernel for CUDA and HIP, boosting Qwen inference speed by up to 20.6%

The new GDN chunked kernel implementation in llama.cpp delivers performance gains ranging from 0.9% to 20.6% across various Qwen models and hardware configurations.

·1·2026.10.14 23:00

ZLUDA Windows Build for AMD GPUs Shared, ~3% Performance Drop

A CUDA-compatible build compiled for Windows based on ZLUDA has been shared, showing approximately 3% slower performance compared to native according to initial evaluations.

·3·2026.09.14 05:00

vLLM Adds HIP W4A16 Kernel Support for AMD

A native HIP W4A16 kernel that optimizes performance on AMD GPUs has been added to vLLM.

·1·2026.05.29 21:00