Michael Mayer 79a3d36334 AI: Drop ROCm support from the llama.cpp install script ... Only Vulkan and CPU (plus Intel SYCL) builds are installed now. Benchmarks show Vulkan generally matches or beats the ROCm build for llama.cpp (except some prompt-processing cases), while ROCm is far more complex to provision: the prebuilt HIP backend needs a matching ROCm runtime, and consumer/iGPU architectures such as gfx1150 additionally require HSA_OVERRIDE_GFX_VERSION because rocBLAS ships no kernels for them. Vulkan drives NVIDIA, AMD, and Intel GPUs on Linux with none of that setup, so auto-detection now selects the Vulkan build for any supported GPU and the CPU build otherwise.		2026-07-10 10:40:58 +02:00
..
dist	AI: Drop ROCm support from the llama.cpp install script	2026-07-10 10:40:58 +02:00
docker
render
setup/nas/raspberry-pi
sql
tools/swaggerfix
build.sh
check-api-request-limits.sh
dav-probe.sh
download-facenet.sh
download-nasnet.sh
download-nsfw.sh
download-scrfd.sh
gettext-clear-fuzzy.sh
gettext-extract.sh
gettext-lint.mjs
gettext-merge.sh
install-qemu.sh
tag-release.sh