From 551d14b11d19b1c8c2917098450f54b8eff49029 Mon Sep 17 00:00:00 2001 From: Donato Capitella Date: Tue, 12 Aug 2025 07:18:35 +0100 Subject: [PATCH] Adding rocm-6.4.3 to README and to refresh script. Adding hipBLASLt. --- README.md | 4 +++- refresh-toolboxes.sh | 2 ++ toolboxes/Dockerfile.rocm-6.4.3 | 6 ++++++ toolboxes/Dockerfile.rocm-6.4.3-rocwmma | 6 ++++++ 4 files changed, 17 insertions(+), 1 deletion(-) diff --git a/README.md b/README.md index a53ccc1..f670a3f 100644 --- a/README.md +++ b/README.md @@ -49,11 +49,13 @@ You can check the containers on DockerHub: https://hub.docker.com/r/kyuz0/amd-st | `vulkan-radv` | Vulkan (Mesa RADV) | Most stable and compatible. Recommended for most users and all models. | | `rocm-6.4.2` | ROCm 6.4.2 (HIP) | Latest stable ROCm. Great for BF16 models. Occasional crashes possible. | | `rocm-6.4.2-rocwmma` | ROCm 6.4.2 (HIP) + ROCWMMA | ROCm with ROCWMMA enabled for improved flash attention on RDNA3+/CDNA. | +| `rocm-6.4.3` | ROCm 6.4.3 (HIP) + hipBLASLt* | Latest stable ROCm. Great for BF16 models. Occasional crashes possible. | +| `rocm-6.4.3-rocwmma` | ROCm 6.4.3 (HIP) + ROCWMMA + hipBLASLt* | ROCm with ROCWMMA enabled for improved flash attention on RDNA3+/CDNA. | | `rocm-7beta` | ROCm 7.0 Beta (HIP) + hipBLASLt* | Latest ROCm beta. No real gain for Llama.cpp. Same model limits as 6.4.2. | | `rocm-7rc` | ROCm 7.0 RC (HIP) + hipBLASLt* | Release candidate for ROCm 7.0. Same behavior as beta. | | `rocm-7rc-rocwmma` | ROCm 7.0 RC (HIP) + ROCWMMA + hipBLASLt* | Release candidate for ROCm 7.0, with hipBLASLt and ROCWMMA for improved flash attention on RDNA3+/CDNA | -* All ROCm 7 toolboxes now export `ROCBLAS_USE_HIPBLASLT=1` as this currently results in better perfromance and stability. +\* All these toolboxes now export `ROCBLAS_USE_HIPBLASLT=1` as this currently results in better perfromance and stability in *MOST* cases. > These containers are **automatically** rebuilt whenever the Llama.cpp master branch is updated, ensuring you get the latest bug fixes and new model support. The easiest way to update to the newest versions is by running the `refresh-toolboxes.sh` [script below](#211-toolbox-refresh-script-automatic-updates). diff --git a/refresh-toolboxes.sh b/refresh-toolboxes.sh index eb21561..c04684f 100644 --- a/refresh-toolboxes.sh +++ b/refresh-toolboxes.sh @@ -9,6 +9,8 @@ TOOLBOXES["llama-vulkan-amdvlk"]="docker.io/kyuz0/amd-strix-halo-toolboxes:vulka TOOLBOXES["llama-vulkan-radv"]="docker.io/kyuz0/amd-strix-halo-toolboxes:vulkan-radv --device /dev/dri --group-add video --security-opt seccomp=unconfined" TOOLBOXES["llama-rocm-6.4.2"]="docker.io/kyuz0/amd-strix-halo-toolboxes:rocm-6.4.2 --device /dev/dri --device /dev/kfd --group-add video --group-add render --group-add sudo --security-opt seccomp=unconfined" TOOLBOXES["llama-rocm-6.4.2-rocwmma"]="docker.io/kyuz0/amd-strix-halo-toolboxes:rocm-6.4.2-rocwmma --device /dev/dri --device /dev/kfd --group-add video --group-add render --group-add sudo --security-opt seccomp=unconfined" +TOOLBOXES["llama-rocm-6.4.3"]="docker.io/kyuz0/amd-strix-halo-toolboxes:rocm-6.4.3 --device /dev/dri --device /dev/kfd --group-add video --group-add render --group-add sudo --security-opt seccomp=unconfined" +TOOLBOXES["llama-rocm-6.4.3-rocwmma"]="docker.io/kyuz0/amd-strix-halo-toolboxes:rocm-6.4.3-rocwmma --device /dev/dri --device /dev/kfd --group-add video --group-add render --group-add sudo --security-opt seccomp=unconfined" TOOLBOXES["llama-rocm-7beta"]="docker.io/kyuz0/amd-strix-halo-toolboxes:rocm-7beta --device /dev/dri --device /dev/kfd --group-add video --group-add render --group-add sudo --security-opt seccomp=unconfined" TOOLBOXES["llama-rocm-7rc"]="docker.io/kyuz0/amd-strix-halo-toolboxes:rocm-7rc --device /dev/dri --device /dev/kfd --group-add video --group-add render --group-add sudo --security-opt seccomp=unconfined" TOOLBOXES["llama-rocm-7rc-rocwmma"]="docker.io/kyuz0/amd-strix-halo-toolboxes:rocm-7rc-rocwmma --device /dev/dri --device /dev/kfd --group-add video --group-add render --group-add sudo --security-opt seccomp=unconfined" diff --git a/toolboxes/Dockerfile.rocm-6.4.3 b/toolboxes/Dockerfile.rocm-6.4.3 index d15f0ed..d163382 100644 --- a/toolboxes/Dockerfile.rocm-6.4.3 +++ b/toolboxes/Dockerfile.rocm-6.4.3 @@ -45,6 +45,12 @@ RUN git clean -xdf \ RUN find /opt/llama.cpp/build -type f -name 'lib*.so*' -exec cp {} /usr/lib64/ \; \ && ldconfig +RUN printf '%s\n' \ + 'export ROCBLAS_USE_HIPBLASLT=1' \ + > /etc/profile.d/rocm.sh \ + && chmod +x /etc/profile.d/rocm.sh \ + && echo 'source /etc/profile.d/rocm.sh' >> /etc/bashrc + COPY gguf-vram-estimator.py /usr/local/bin/gguf-vram-estimator.py RUN chmod +x /usr/local/bin/gguf-vram-estimator.py diff --git a/toolboxes/Dockerfile.rocm-6.4.3-rocwmma b/toolboxes/Dockerfile.rocm-6.4.3-rocwmma index 0e53987..15512a1 100644 --- a/toolboxes/Dockerfile.rocm-6.4.3-rocwmma +++ b/toolboxes/Dockerfile.rocm-6.4.3-rocwmma @@ -55,5 +55,11 @@ RUN find /opt/llama.cpp/build -type f -name 'lib*.so*' -exec cp {} /usr/lib64/ \ COPY gguf-vram-estimator.py /usr/local/bin/gguf-vram-estimator.py RUN chmod +x /usr/local/bin/gguf-vram-estimator.py +RUN printf '%s\n' \ + 'export ROCBLAS_USE_HIPBLASLT=1' \ + > /etc/profile.d/rocm.sh \ + && chmod +x /etc/profile.d/rocm.sh \ + && echo 'source /etc/profile.d/rocm.sh' >> /etc/bashrc + # Default to interactive shell CMD ["/bin/bash"]