diff --git a/.github/labeler.yml b/.github/labeler.yml
index 89436740d..1b47bc968 100644
--- a/.github/labeler.yml
+++ b/.github/labeler.yml
@@ -3,19 +3,18 @@ Kompute:
     - changed-files:
         - any-glob-to-any-file:
             - ggml/include/ggml-kompute.h
-            - ggml/src/ggml-kompute.cpp
+            - ggml/src/ggml-kompute/**
             - README-kompute.md
 Apple Metal:
     - changed-files:
         - any-glob-to-any-file:
             - ggml/include/ggml-metal.h
-            - ggml/src/ggml-metal.cpp
+            - ggml/src/ggml-metal/**
             - README-metal.md
 SYCL:
     - changed-files:
         - any-glob-to-any-file:
             - ggml/include/ggml-sycl.h
-            - ggml/src/ggml-sycl.cpp
             - ggml/src/ggml-sycl/**
             - docs/backend/SYCL.md
             - examples/sycl/**
@@ -27,8 +26,8 @@ Nvidia GPU:
 Vulkan:
     - changed-files:
         - any-glob-to-any-file:
-            - ggml/ggml_vk_generate_shaders.py
-            - ggml/src/ggml-vulkan*
+            - ggml/include/ggml-vulkan.h
+            - ggml/src/ggml-vulkan/**
 documentation:
     - changed-files:
         - any-glob-to-any-file:
@@ -75,11 +74,7 @@ server:
 ggml:
     - changed-files:
         - any-glob-to-any-file:
-            - ggml/include/ggml*.h
-            - ggml/src/ggml*.c
-            - ggml/src/ggml*.cpp
-            - ggml/src/ggml*.h
-            - ggml-cuda/**
+            - ggml/**
 nix:
     - changed-files:
         - any-glob-to-any-file:
diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml
index abaf2c504..6281663ec 100644
--- a/.github/workflows/build.yml
+++ b/.github/workflows/build.yml
@@ -871,8 +871,65 @@ jobs:
           path: llama-${{ steps.tag.outputs.name }}-bin-win-${{ matrix.build }}.zip
           name: llama-bin-win-${{ matrix.build }}.zip
 
+  ubuntu-latest-cmake-cuda:
+    runs-on: ubuntu-latest
+    container: nvidia/cuda:12.6.2-devel-ubuntu24.04
+
+    steps:
+        - name: Clone
+          id: checkout
+          uses: actions/checkout@v4
+
+        - name: Install dependencies
+          env:
+            DEBIAN_FRONTEND: noninteractive
+          run: |
+              apt update
+              apt install -y cmake build-essential ninja-build libgomp1 git
+
+        - name: Build with CMake
+          run: |
+            cmake -S . -B build -G Ninja -DCMAKE_BUILD_TYPE=Release -DGGML_NATIVE=OFF -DGGML_CUDA=ON -DCMAKE_CUDA_ARCHITECTURES=89-real -DCMAKE_EXE_LINKER_FLAGS=-Wl,--allow-shlib-undefined -DLLAMA_FATAL_WARNINGS=ON
+            cmake --build build
+
   windows-latest-cmake-cuda:
+    runs-on: windows-latest
+
+    strategy:
+      matrix:
+        cuda: ['12.6.2']
+        build: ['cuda']
+
+    steps:
+      - name: Clone
+        id: checkout
+        uses: actions/checkout@v4
+
+      - name: Install CUDA toolkit
+        id: cuda-toolkit
+        uses: Jimver/cuda-toolkit@v0.2.19
+        with:
+          cuda: ${{ matrix.cuda }}
+          method: 'network'
+          sub-packages: '["nvcc", "cudart", "cublas", "cublas_dev", "thrust", "visual_studio_integration"]'
+
+      - name: Install Ninja
+        id: install_ninja
+        run: |
+          choco install ninja
+
+      - name: Build
+        id: cmake_build
+        shell: cmd
+        run: |
+          call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvars64.bat"
+          cmake -S . -B build -G "Ninja Multi-Config" -DGGML_NATIVE=OFF -DGGML_CUDA=ON -DBUILD_SHARED_LIBS=ON -DGGML_RPC=ON -DCMAKE_CUDA_ARCHITECTURES=89-real
+          cmake --build build --config Release -t ggml-cuda
+          cmake --build build --config Release
+
+  windows-2019-cmake-cuda:
     runs-on: windows-2019
+    if: ${{ github.event == 'push' && github.ref == 'refs/heads/master' }}
 
     strategy:
       matrix:
@@ -1173,7 +1230,7 @@ jobs:
       - macOS-latest-make
       - macOS-latest-cmake
       - windows-latest-cmake
-      - windows-latest-cmake-cuda
+      - windows-2019-cmake-cuda
       - windows-latest-cmake-hip-release
       - macOS-latest-cmake-arm64
       - macOS-latest-cmake-x64
diff --git a/.github/workflows/nix-ci.yml b/.github/workflows/nix-ci.yml
index 8ecbbe53b..3fe941576 100644
--- a/.github/workflows/nix-ci.yml
+++ b/.github/workflows/nix-ci.yml
@@ -5,8 +5,10 @@ on:
   push:
     branches:
       - master
+    paths: ['.github/workflows/nix-ci.yml', '**/flake.nix', '**/flake.lock', '**/CMakeLists.txt', '**/Makefile', '**/*.h', '**/*.hpp', '**/*.c', '**/*.cpp', '**/*.cu', '**/*.cuh', '**/*.swift', '**/*.m', '**/*.metal']
   pull_request:
     types: [opened, synchronize, reopened]
+    paths: ['.github/workflows/nix-ci.yml', '**/flake.nix', '**/flake.lock', '**/CMakeLists.txt', '**/Makefile', '**/*.h', '**/*.hpp', '**/*.c', '**/*.cpp', '**/*.cu', '**/*.cuh', '**/*.swift', '**/*.m', '**/*.metal']
 
 concurrency:
   group: ${{ github.workflow }}-${{ github.head_ref && github.ref || github.run_id }}
diff --git a/.github/workflows/python-lint.yml b/.github/workflows/python-lint.yml
index a8d46f31d..ddfdf73b8 100644
--- a/.github/workflows/python-lint.yml
+++ b/.github/workflows/python-lint.yml
@@ -1,6 +1,13 @@
 name: flake8 Lint
 
-on: [push, pull_request]
+on:
+  push:
+    branches:
+      - master
+    paths: ['.github/workflows/python-lint.yml', '**/*.py']
+  pull_request:
+    types: [opened, synchronize, reopened]
+    paths: ['.github/workflows/python-lint.yml', '**/*.py']
 
 concurrency:
   group: ${{ github.workflow }}-${{ github.head_ref && github.ref || github.run_id }}