chore(cuda): use newer ami (#2131)

Author: gaxiom

.github/runs-on.yml

@@ -5,20 +5,12 @@ images:
     owner: "405894839083"
     name: "arm64-rust-riscv-dev-32gb"
     ami: "ami-040a96edd406c9478"
-  # RunsOn nvidia image with cuda 12.9 installed (latest is cuda 13)
-  # See: https://github.com/runs-on/runner-images-for-aws
-  ubuntu24-gpu-x64-cuda12-9:
-    platform: "linux"
-    arch: "x86_64"
-    owner: "135269210855"
-    name: "runs-on-v2.2-ubuntu24-gpu-x64-20250731145109"
-    ami: "ami-017d4d69faf854c81"
 
 runners:
   # runners for running tests
   test-gpu-nvidia:
     family: ["g6", "g5", "g4dn", "g6e"]
-    image: ubuntu24-gpu-x64-cuda12-9
+    image: ubuntu24-gpu-x64
     cpu: [4, 8, 16, 32]
     disk: default
     spot: capacity-optimized

.github/workflows/base-tests.cuda.yml

@@ -27,7 +27,7 @@ env:
 jobs:
   tests-cuda:
     runs-on:
-      - runs-on=${{ github.run_id }}-base-tests-cuda-${{ github.run_attempt }}/family=g6+g6e/image=ubuntu24-gpu-x64-cuda12-9/extras=s3-cache
+      - runs-on=${{ github.run_id }}-base-tests-cuda-${{ github.run_attempt }}/family=g6+g6e/image=ubuntu24-gpu-x64/extras=s3-cache
 
     steps:
       - uses: runs-on/action@v2

.github/workflows/benchmark-call.yml

@@ -128,7 +128,7 @@ jobs:
   bench-new:
     name: Run benchmark on workflow ref/branch
     runs-on:
-      - runs-on=${{ github.run_id }}-benchcall-${{ github.run_attempt }}-${{ inputs.job_index || inputs.benchmark_id || inputs.benchmark_name || github.event.inputs.benchmark_name }}/family=${{ inputs.instance_type || github.event.inputs.instance_type }}/image=${{ startsWith(inputs.instance_type || github.event.inputs.instance_type, 'g') && 'ubuntu24-gpu-x64-cuda12-9' || contains(inputs.instance_type || github.event.inputs.instance_type, 'g.') && 'ubuntu24-full-arm64' || 'ubuntu24-full-x64' }}/spot=capacity-optimized/extras=s3-cache
+      - runs-on=${{ github.run_id }}-benchcall-${{ github.run_attempt }}-${{ inputs.job_index || inputs.benchmark_id || inputs.benchmark_name || github.event.inputs.benchmark_name }}/family=${{ inputs.instance_type || github.event.inputs.instance_type }}/image=${{ startsWith(inputs.instance_type || github.event.inputs.instance_type, 'g') && 'ubuntu24-gpu-x64' || contains(inputs.instance_type || github.event.inputs.instance_type, 'g.') && 'ubuntu24-full-arm64' || 'ubuntu24-full-x64' }}/spot=capacity-optimized/extras=s3-cache
     steps:
       - uses: runs-on/action@v2
       ##########################################################################

.github/workflows/lints.yml

@@ -123,6 +123,9 @@ jobs:
 
       - name: Run clang-tidy
         run: |
+          set -eux
+          wget -qO- https://apt.llvm.org/llvm-snapshot.gpg.key | sudo tee /etc/apt/trusted.gpg.d/apt.llvm.org.asc >/dev/null
+          echo "deb http://apt.llvm.org/noble/ llvm-toolchain-noble main" | sudo tee /etc/apt/sources.list.d/llvm-toolchain.list
           sudo apt-get update
           sudo apt-get install -y clang-tidy
           ./scripts/run_clang_tidy.sh

.github/workflows/riscv.yml

@@ -31,7 +31,7 @@ jobs:
       matrix:
         platform:
           - { runner: "64cpu-linux-arm64", image: "arm64-rust-riscv-dev-32gb" }
-          - { runner: "test-gpu-nvidia", image: "ubuntu24-gpu-x64-cuda12-9" }
+          - { runner: "test-gpu-nvidia", image: "ubuntu24-gpu-x64" }
 
     runs-on:
       - runs-on=${{ github.run_id }}-riscv-${{ github.run_attempt }}-${{ strategy.job-index }}/runner=${{ matrix.platform.runner }}/image=${{ matrix.platform.image }}/extras=s3-cache