feat: upgrade ROCm toolboxes to 7.2.2 and update documentation and CI configurations

.github/workflows/build_and_publish.yml

@@ -28,7 +28,7 @@ jobs:
           IN='${{ inputs.backends }}'
 
           if [[ "$IN" == "all" || -z "$IN" ]]; then
-            JSON='["rocm-6.4.4","rocm-7.2.1","rocm-7.2.1-pr21344","rocm7-nightlies","vulkan-amdvlk","vulkan-radv"]'
+            JSON='["rocm-6.4.4","rocm-7.2.2","rocm-7.2.2-pr21344","rocm7-nightlies","vulkan-amdvlk","vulkan-radv"]'
           else
             # Remove spaces and build JSON array from comma list
             IN_CLEAN=$(echo "$IN" | tr -d '[:space:]')

.github/workflows/prune-old-toolboxes.yml

@@ -44,7 +44,7 @@ jobs:
         run: |
           IN='${{ github.event.inputs.backends }}'
           if [[ "$IN" == "all" || -z "$IN" ]]; then
-            JSON='["rocm-6.4.2","rocm-6.4.3","rocm-6.4.4","rocm-7.1.1","rocm-7.2","rocm-7.2.1","rocm-7.2.1-pr21344","rocm-7beta","rocm7-nightlies","vulkan-amdvlk","vulkan-radv"]'
+            JSON='["rocm-6.4.2","rocm-6.4.3","rocm-6.4.4","rocm-7.1.1","rocm-7.2","rocm-7.2.1","rocm-7.2.1-pr21344","rocm-7.2.2","rocm-7.2.2-pr21344","rocm-7beta","rocm7-nightlies","vulkan-amdvlk","vulkan-radv"]'
           else
             IN_CLEAN=$(echo "$IN" | tr -d '[:space:]')
             JSON='["'${IN_CLEAN//,/\",\"}'"]'

AGENTS.md

@@ -8,7 +8,7 @@
 *   **Hardware / Drivers**: AMD "Strix Halo" APUs (Gfx1151). Implementations depend on ROCm (v6.4.4, v7.x) and Vulkan (Mesa RADV, AMDVLK).
 
 ## Repository Structure Overview
-*   `/toolboxes/`: Dockerfiles used to build the container images (e.g., `rocm-6.4.4`, `rocm-7.2.1`, `vulkan-radv`). These often use multi-stage builds to compile Llama.cpp and extract standalone binaries.
+*   `/toolboxes/`: Dockerfiles used to build the container images (e.g., `rocm-6.4.4`, `rocm-7.2.2`, `vulkan-radv`). These often use multi-stage builds to compile Llama.cpp and extract standalone binaries.
 *   `/benchmark/`: Shell scripts and Python utilities (like `generate_results_json.py`) to systematically test Llama.cpp throughput, latency, and RPC performance.
 *   `/docs/`: Markdown documents, along with HTML/CSS/JS (e.g., `index.html`, `assets/`) for the GitHub Pages website (`strix-halo-toolboxes.com`), plus interactive benchmark viewers and documentation on VRAM estimation.
 *   `/scripts/`: Python utilities, including `run_distributed_llama.py` for distributed inference across nodes.

README.md

@@ -34,24 +34,18 @@ This is a hobby project maintained in my spare time. If you find these toolboxes
 ## Stable Configuration
 
 - **OS**: Fedora 42/43
-- **Linux Kernel**: 6.18.6-200
+- **Linux Kernel**: 6.18.9-200.fc43.x86_64
 - **Linux Firmware**: 20260110
 
 This is currently the most stable setup. Kernels older than 6.18.4 have a bug that causes stability issues on gfx1151 and should be avoided. Also, **do NOT use `linux-firmware-20251125`.** It breaks ROCm support on Strix Halo (instability/crashes).
 
 > ⚠️ **Important**: See [Host Configuration](#host-configuration) for critical kernel parameters.
 
-## ROCm 7 Performance Regression Workaround
-
-The performance regression previously observed in ROCm 7+ builds (compared to ROCm 6.4.4) has been **resolved in the toolboxes** via a workaround.
-
-The issue was caused by a compiler regression (llvm/llvm-project#147700) affecting loop unrolling thresholds. We have applied the workaround (`-mllvm --amdgpu-unroll-threshold-local=600`) in the latest toolbox builds, restoring full performance.
-
-This workaround will be removed once the upstream fix lands. For details, see the issue: [kyuz0/amd-strix-halo-toolboxes#45](https://github.com/kyuz0/amd-strix-halo-toolboxes/issues/45)
-
-
 ## Supported Toolboxes
 
+> [!WARNING]
+> Current `rocm7-nightlies` builds have a bug that caps memory allocation to 64GB. If you need larger models, prefer stable builds like `rocm-7.2.2` (performance is similar). Track the issue here: https://github.com/ROCm/TheRock/issues/4645
+
 You can check the containers on DockerHub: [kyuz0/amd-strix-halo-toolboxes](https://hub.docker.com/r/kyuz0/amd-strix-halo-toolboxes/tags).
 
 | Container Tag | Backend/Stack | Purpose / Notes |
@@ -59,7 +53,7 @@ You can check the containers on DockerHub: [kyuz0/amd-strix-halo-toolboxes](http
 | `vulkan-amdvlk` | Vulkan (AMDVLK) | Fastest backend—AMD open-source driver. ≤2 GiB single buffer allocation limit, some large models won't load. |
 | `vulkan-radv` | Vulkan (Mesa RADV) | Most stable and compatible. Recommended for most users and all models. |
 | `rocm-6.4.4` | ROCm 6.4.4 (Fedora 43) | Latest stable 6.x build. Uses Fedora 43 packages with backported patch for **kernel 6.18.4+** support. |
-| `rocm-7.2.1` | ROCm 7.2.1 | Latest stable 7.x build. Includes patch for **kernel 6.18.4+** support. |
+| `rocm-7.2.2` | ROCm 7.2.2 | Latest stable 7.x build. Includes patch for **kernel 6.18.4+** support. |
 | `rocm7-nightlies` | ROCm 7 Nightly | Tracks nightly builds. Includes patch for **kernel 6.18.4+** support. |
 
 > These containers are **automatically** rebuilt whenever the Llama.cpp master branch is updated. Legacy images (`rocm-6.4.2`, `rocm-6.4.3`, `rocm-7.1.1`) are excluded from this list.
@@ -79,12 +73,12 @@ toolbox enter llama-vulkan-radv
 
 **Option B: ROCm (Recommended for Performance)**
 ```sh
-toolbox create llama-rocm-7.2.1 \
+toolbox create llama-rocm-7.2.2 \
-  --image docker.io/kyuz0/amd-strix-halo-toolboxes:rocm-7.2.1 \
+  --image docker.io/kyuz0/amd-strix-halo-toolboxes:rocm-7.2.2 \
   -- --device /dev/dri --device /dev/kfd \
   --group-add video --group-add render --group-add sudo --security-opt seccomp=unconfined
 
-toolbox enter llama-rocm-7.2
+toolbox enter llama-rocm-7.2.2
 ```
 
 ### 2. Check GPU Access

benchmark/run_benchmarks.sh

@@ -62,8 +62,8 @@ echo
 
 declare -A CMDS=(
   [rocm6_4_4]="toolbox run -c llama-rocm-6.4.4 -- /usr/local/bin/llama-bench"
-  [rocm-7_2_1]="toolbox run -c llama-rocm-7.2.1 -- /usr/local/bin/llama-bench"
+  [rocm-7_2_2]="toolbox run -c llama-rocm-7.2.2 -- /usr/local/bin/llama-bench"
-  [rocm-7_2_1-pr21344]="toolbox run -c llama-rocm-7.2.1-pr21344 -- /usr/local/bin/llama-bench"
+  [rocm-7_2_2-pr21344]="toolbox run -c llama-rocm-7.2.2-pr21344 -- /usr/local/bin/llama-bench"
   [rocm7-nightlies]="toolbox run -c llama-rocm7-nightlies -- /usr/local/bin/llama-bench"
   [vulkan_amdvlk]="toolbox run -c llama-vulkan-amdvlk -- /usr/sbin/llama-bench"
   [vulkan_radv]="toolbox run -c llama-vulkan-radv -- /usr/sbin/llama-bench"

refresh-toolboxes.sh

@@ -8,8 +8,8 @@ declare -A TOOLBOXES
 TOOLBOXES["llama-vulkan-amdvlk"]="docker.io/kyuz0/amd-strix-halo-toolboxes:vulkan-amdvlk --device /dev/dri --group-add video --security-opt seccomp=unconfined"
 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-6.4.4"]="docker.io/kyuz0/amd-strix-halo-toolboxes:rocm-6.4.4 --device /dev/dri --device /dev/kfd --group-add video --group-add render --group-add sudo --security-opt seccomp=unconfined"
-TOOLBOXES["llama-rocm-7.2.1"]="docker.io/kyuz0/amd-strix-halo-toolboxes:rocm-7.2.1 --device /dev/dri --device /dev/kfd --group-add video --group-add render --group-add sudo --security-opt seccomp=unconfined"
+TOOLBOXES["llama-rocm-7.2.2"]="docker.io/kyuz0/amd-strix-halo-toolboxes:rocm-7.2.2 --device /dev/dri --device /dev/kfd --group-add video --group-add render --group-add sudo --security-opt seccomp=unconfined"
-TOOLBOXES["llama-rocm-7.2.1-pr21344"]="docker.io/kyuz0/amd-strix-halo-toolboxes:rocm-7.2.1-pr21344 --device /dev/dri --device /dev/kfd --group-add video --group-add render --group-add sudo --security-opt seccomp=unconfined"
+TOOLBOXES["llama-rocm-7.2.2-pr21344"]="docker.io/kyuz0/amd-strix-halo-toolboxes:rocm-7.2.2-pr21344 --device /dev/dri --device /dev/kfd --group-add video --group-add render --group-add sudo --security-opt seccomp=unconfined"
 TOOLBOXES["llama-rocm7-nightlies"]="docker.io/kyuz0/amd-strix-halo-toolboxes:rocm7-nightlies --device /dev/dri --device /dev/kfd --group-add video --group-add render --group-add sudo --security-opt seccomp=unconfined"
 
 function usage() {

toolboxes/Dockerfile.rocm-7.2.2

@@ -1,12 +1,12 @@
 # build stage
 FROM registry.fedoraproject.org/fedora:43 AS builder
 
-# rocm 7.2.1 repo
+# rocm 7.2.2 repo
 RUN <<'EOF'
 tee /etc/yum.repos.d/rocm.repo <<REPO
-[ROCm-7.2.1]
+[ROCm-7.2.2]
-name=ROCm7.2.1
+name=ROCm7.2.2
-baseurl=https://repo.radeon.com/rocm/rhel10/7.2.1/main
+baseurl=https://repo.radeon.com/rocm/rhel10/7.2.2/main
 enabled=1
 priority=50
 gpgcheck=1
@@ -69,12 +69,12 @@ RUN chmod +x /usr/local/bin/gguf-vram-estimator.py
 # runtime stage
 FROM registry.fedoraproject.org/fedora-minimal:43
 
-# rocm 7.2.1 repo
+# rocm 7.2.2 repo
 RUN <<'EOF'
 tee /etc/yum.repos.d/rocm.repo <<REPO
-[ROCm-7.2.1]
+[ROCm-7.2.2]
-name=ROCm7.2.1
+name=ROCm7.2.2
-baseurl=https://repo.radeon.com/rocm/rhel10/7.2.1/main
+baseurl=https://repo.radeon.com/rocm/rhel10/7.2.2/main
 enabled=1
 priority=50
 gpgcheck=1

toolboxes/Dockerfile.rocm-7.2.2-pr21344

@@ -1,14 +1,14 @@
 # build stage
-# Based on Dockerfile.rocm-7.2.1, but clones pedapudi/llama.cpp@gfx1151-opt
+# Based on Dockerfile.rocm-7.2.2, but clones pedapudi/llama.cpp@gfx1151-opt
 # (PR #21344: gfx1151 nwarps, tile sizing to curb VGPR pressure)
 FROM registry.fedoraproject.org/fedora:43 AS builder
 
-# rocm 7.2.1 repo
+# rocm 7.2.2 repo
 RUN <<'EOF'
 tee /etc/yum.repos.d/rocm.repo <<REPO
-[ROCm-7.2.1]
+[ROCm-7.2.2]
-name=ROCm7.2.1
+name=ROCm7.2.2
-baseurl=https://repo.radeon.com/rocm/rhel10/7.2.1/main
+baseurl=https://repo.radeon.com/rocm/rhel10/7.2.2/main
 enabled=1
 priority=50
 gpgcheck=1
@@ -78,12 +78,12 @@ RUN chmod +x /usr/local/bin/gguf-vram-estimator.py
 # runtime stage
 FROM registry.fedoraproject.org/fedora-minimal:43
 
-# rocm 7.2.1 repo
+# rocm 7.2.2 repo
 RUN <<'EOF'
 tee /etc/yum.repos.d/rocm.repo <<REPO
-[ROCm-7.2.1]
+[ROCm-7.2.2]
-name=ROCm7.2.1
+name=ROCm7.2.2
-baseurl=https://repo.radeon.com/rocm/rhel10/7.2.1/main
+baseurl=https://repo.radeon.com/rocm/rhel10/7.2.2/main
 enabled=1
 priority=50
 gpgcheck=1