Product
InferBench
A benchmarking tool for local LLM inference performance on AMD GPUs, focused on comparing the inference performance of the Vulkan backend against the HIP (ROCm) backend on the same hardware, providing users with systematic empirical measurement data.
Timeline (last 90 days)
Sep 20
The ROCmFix and InferBench projects were shared on Hacker News.
Expired50%
Sep 20
InferBench focuses on comparing inference performance between the Vulkan backend and the HIP backend on AMD hardware.
Unverified50%