From 51e46d0b48d6f1d3ef5ddd882dddc7197b48ac14 Mon Sep 17 00:00:00 2001
From: nihuini <nihuini@tencent.com>
Date: Mon, 1 Apr 2024 16:45:30 +0800
Subject: [PATCH 01/19] add visionos ci

---
 .github/workflows/visionos-cpu.yml | 73 ++++++++++++++++++++++++++++++
 1 file changed, 73 insertions(+)
 create mode 100644 .github/workflows/visionos-cpu.yml

diff --git a/.github/workflows/visionos-cpu.yml b/.github/workflows/visionos-cpu.yml
new file mode 100644
index 000000000000..ac82a4789ccc
--- /dev/null
+++ b/.github/workflows/visionos-cpu.yml
@@ -0,0 +1,73 @@
+name: visionos-cpu
+on:
+  push:
+    branches: [master]
+    paths:
+    - '.github/workflows/visionos-cpu.yml'
+    - 'toolchains/ios.toolchain.cmake'
+    - 'CMakeLists.txt'
+    - 'cmake/**'
+    - 'src/*'
+    - 'src/layer/*'
+    - 'src/layer/arm/**'
+  pull_request:
+    branches: [master]
+    paths:
+    - '.github/workflows/visionos-cpu.yml'
+    - 'toolchains/ios.toolchain.cmake'
+    - 'CMakeLists.txt'
+    - 'cmake/**'
+    - 'src/*'
+    - 'src/layer/*'
+    - 'src/layer/arm/**'
+concurrency:
+  group: visionos-cpu-${{ github.ref }}
+  cancel-in-progress: true
+env:
+  DEVELOPER_DIR: /Applications/Xcode_15.2.app/Contents/Developer
+  VISIONOS_DEPLOYMENT_TARGET: '1.0'
+  ENABLE_BITCODE: OFF
+  ENABLE_ARC: OFF
+  ENABLE_VISIBILITY: OFF
+permissions:
+  contents: read
+
+jobs:
+  visionos:
+    runs-on: macos-13
+    steps:
+    - uses: actions/checkout@v4
+
+    - name: build-arm64
+      run: |
+        mkdir build-arm64 && cd build-arm64
+        cmake -DCMAKE_TOOLCHAIN_FILE=../toolchains/ios.toolchain.cmake -DPLATFORM=VISIONOS -DARCHS="arm64" \
+            -DDEPLOYMENT_TARGET=$VISIONOS_DEPLOYMENT_TARGET -DENABLE_BITCODE=$ENABLE_BITCODE -DENABLE_ARC=$ENABLE_ARC -DENABLE_VISIBILITY=$ENABLE_VISIBILITY \
+            -DNCNN_OPENMP=OFF -DNCNN_THREADS=OFF \
+            -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=install \
+            ..
+        cmake --build . -j 3
+
+  visionos-simulator:
+    runs-on: macos-13
+    steps:
+    - uses: actions/checkout@v4
+
+    - name: build-x86_64
+      run: |
+        mkdir build-x86_64 && cd build-x86_64
+        cmake -DCMAKE_TOOLCHAIN_FILE=../toolchains/ios.toolchain.cmake -DPLATFORM=SIMULATOR_VISIONOS -DARCHS="x86_64" \
+            -DDEPLOYMENT_TARGET=$VISIONOS_DEPLOYMENT_TARGET -DENABLE_BITCODE=$ENABLE_BITCODE -DENABLE_ARC=$ENABLE_ARC -DENABLE_VISIBILITY=$ENABLE_VISIBILITY \
+            -DNCNN_OPENMP=OFF -DNCNN_THREADS=OFF \
+            -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=install \
+            ..
+        cmake --build . -j 3
+    - name: build-arm64
+      run: |
+        mkdir build-arm64 && cd build-arm64
+        cmake -DCMAKE_TOOLCHAIN_FILE=../toolchains/ios.toolchain.cmake -DPLATFORM=SIMULATOR_VISIONOS -DARCHS="arm64" \
+            -DDEPLOYMENT_TARGET=$VISIONOS_DEPLOYMENT_TARGET -DENABLE_BITCODE=$ENABLE_BITCODE -DENABLE_ARC=$ENABLE_ARC -DENABLE_VISIBILITY=$ENABLE_VISIBILITY \
+            -DNCNN_OPENMP=OFF -DNCNN_THREADS=OFF \
+            -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=install \
+            ..
+        cmake --build . -j 3

From 8f963b84def89821ca7fd906ecff66991b6ef94d Mon Sep 17 00:00:00 2001
From: nihuini <nihuini@tencent.com>
Date: Mon, 1 Apr 2024 17:07:25 +0800
Subject: [PATCH 02/19] update watchos ci

---
 .github/workflows/watchos-cpu.yml | 91 -------------------------------
 .github/workflows/watchos.yml     | 79 +++++++++++++++++++++++++++
 2 files changed, 79 insertions(+), 91 deletions(-)
 delete mode 100644 .github/workflows/watchos-cpu.yml
 create mode 100644 .github/workflows/watchos.yml

diff --git a/.github/workflows/watchos-cpu.yml b/.github/workflows/watchos-cpu.yml
deleted file mode 100644
index adcea8103750..000000000000
--- a/.github/workflows/watchos-cpu.yml
+++ /dev/null
@@ -1,91 +0,0 @@
-name: watchos-cpu
-on:
-  push:
-    branches: [master]
-    paths:
-    - '.github/workflows/watchos-cpu.yml'
-    - 'toolchains/ios.toolchain.cmake'
-    - 'CMakeLists.txt'
-    - 'cmake/**'
-    - 'src/*'
-    - 'src/layer/*'
-    - 'src/layer/arm/**'
-  pull_request:
-    branches: [master]
-    paths:
-    - '.github/workflows/watchos-cpu.yml'
-    - 'toolchains/ios.toolchain.cmake'
-    - 'CMakeLists.txt'
-    - 'cmake/**'
-    - 'src/*'
-    - 'src/layer/*'
-    - 'src/layer/arm/**'
-concurrency:
-  group: watchos-cpu-${{ github.ref }}
-  cancel-in-progress: true
-env:
-  DEVELOPER_DIR: /Applications/Xcode_13.4.1.app/Contents/Developer
-  WATCHOS_DEPLOYMENT_TARGET: '2.0'
-  ENABLE_BITCODE: OFF
-  ENABLE_ARC: OFF
-  ENABLE_VISIBILITY: OFF
-permissions:
-  contents: read
-
-jobs:
-  watchos:
-    runs-on: macos-12
-    steps:
-    - uses: actions/checkout@v4
-
-    - name: build-armv7k
-      run: |
-        mkdir build-armv7k && cd build-armv7k
-        cmake -DCMAKE_TOOLCHAIN_FILE=../toolchains/ios.toolchain.cmake -DPLATFORM=WATCHOS -DARCHS="armv7k" \
-            -DDEPLOYMENT_TARGET=$WATCHOS_DEPLOYMENT_TARGET -DENABLE_BITCODE=$ENABLE_BITCODE -DENABLE_ARC=$ENABLE_ARC -DENABLE_VISIBILITY=$ENABLE_VISIBILITY \
-            -DNCNN_OPENMP=OFF -DNCNN_THREADS=OFF \
-            -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=install \
-            ..
-        cmake --build . -j 3
-    - name: build-arm64_32
-      run: |
-        mkdir build-arm64_32 && cd build-arm64_32
-        cmake -DCMAKE_TOOLCHAIN_FILE=../toolchains/ios.toolchain.cmake -DPLATFORM=WATCHOS -DARCHS="arm64_32" \
-            -DDEPLOYMENT_TARGET=$WATCHOS_DEPLOYMENT_TARGET -DENABLE_BITCODE=$ENABLE_BITCODE -DENABLE_ARC=$ENABLE_ARC -DENABLE_VISIBILITY=$ENABLE_VISIBILITY \
-            -DNCNN_OPENMP=OFF -DNCNN_THREADS=OFF \
-            -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=install \
-            ..
-        cmake --build . -j 3
-
-  watchos-simulator:
-    runs-on: macos-12
-    steps:
-    - uses: actions/checkout@v4
-
-    - name: build-i386
-      run: |
-        mkdir build-i386 && cd build-i386
-        cmake -DCMAKE_TOOLCHAIN_FILE=../toolchains/ios.toolchain.cmake -DPLATFORM=SIMULATOR_WATCHOS -DARCHS="i386" \
-            -DDEPLOYMENT_TARGET=$WATCHOS_DEPLOYMENT_TARGET -DENABLE_BITCODE=$ENABLE_BITCODE -DENABLE_ARC=$ENABLE_ARC -DENABLE_VISIBILITY=$ENABLE_VISIBILITY \
-            -DNCNN_OPENMP=OFF -DNCNN_THREADS=OFF \
-            -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=install \
-            ..
-        cmake --build . -j 3
-    - name: build-x86_64
-      run: |
-        mkdir build-x86_64 && cd build-x86_64
-        cmake -DCMAKE_TOOLCHAIN_FILE=../toolchains/ios.toolchain.cmake -DPLATFORM=SIMULATOR_WATCHOS -DARCHS="x86_64" \
-            -DDEPLOYMENT_TARGET=$WATCHOS_DEPLOYMENT_TARGET -DENABLE_BITCODE=$ENABLE_BITCODE -DENABLE_ARC=$ENABLE_ARC -DENABLE_VISIBILITY=$ENABLE_VISIBILITY \
-            -DNCNN_OPENMP=OFF -DNCNN_THREADS=OFF \
-            -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=install \
-            ..
-        cmake --build . -j 3
-    - name: build-arm64
-      run: |
-        mkdir build-arm64 && cd build-arm64
-        cmake -DCMAKE_TOOLCHAIN_FILE=../toolchains/ios.toolchain.cmake -DPLATFORM=SIMULATOR_WATCHOS -DARCHS="arm64" \
-            -DDEPLOYMENT_TARGET=$WATCHOS_DEPLOYMENT_TARGET -DENABLE_BITCODE=$ENABLE_BITCODE -DENABLE_ARC=$ENABLE_ARC -DENABLE_VISIBILITY=$ENABLE_VISIBILITY \
-            -DNCNN_OPENMP=OFF -DNCNN_THREADS=OFF \
-            -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=install \
-            ..
-        cmake --build . -j 3
diff --git a/.github/workflows/watchos.yml b/.github/workflows/watchos.yml
new file mode 100644
index 000000000000..2ff987891249
--- /dev/null
+++ b/.github/workflows/watchos.yml
@@ -0,0 +1,79 @@
+name: watchos
+on:
+  push:
+    branches: [master]
+    paths:
+    - '.github/workflows/watchos.yml'
+    - 'toolchains/ios.toolchain.cmake'
+    - 'CMakeLists.txt'
+    - 'cmake/**'
+    - 'src/*'
+    - 'src/layer/*'
+    - 'src/layer/arm/**'
+    - 'src/layer/x86/**'
+  pull_request:
+    branches: [master]
+    paths:
+    - '.github/workflows/watchos.yml'
+    - 'toolchains/ios.toolchain.cmake'
+    - 'CMakeLists.txt'
+    - 'cmake/**'
+    - 'src/*'
+    - 'src/layer/*'
+    - 'src/layer/arm/**'
+    - 'src/layer/x86/**'
+concurrency:
+  group: watchos-${{ github.ref }}
+  cancel-in-progress: true
+env:
+  DEVELOPER_DIR: /Applications/Xcode_14.3.1.app/Contents/Developer
+  WATCHOS_DEPLOYMENT_TARGET: '2.0'
+  ENABLE_BITCODE: OFF
+  ENABLE_ARC: OFF
+  ENABLE_VISIBILITY: OFF
+permissions:
+  contents: read
+
+jobs:
+  build:
+    runs-on: macos-13
+    env:
+      COMMON_CMAKE_OPTIONS: |
+        -DCMAKE_TOOLCHAIN_FILE=../toolchains/ios.toolchain.cmake \
+        -DDEPLOYMENT_TARGET=$WATCHOS_DEPLOYMENT_TARGET \
+        -DENABLE_BITCODE=$ENABLE_BITCODE \
+        -DENABLE_ARC=$ENABLE_ARC \
+        -DENABLE_VISIBILITY=$ENABLE_VISIBILITY \
+        -DNCNN_OPENMP=OFF -DNCNN_THREADS=OFF \
+        -DCMAKE_INSTALL_PREFIX=install \
+        -DCMAKE_BUILD_TYPE=Release \
+
+    steps:
+    - uses: actions/checkout@v4
+
+    - name: armv7k
+      run: |
+        mkdir build-armv7k && cd build-armv7k
+        cmake ${{ env.COMMON_CMAKE_OPTIONS }} -DPLATFORM=WATCHOS -DARCHS="armv7k" ..
+        cmake --build . -j 4
+    - name: arm64_32
+      run: |
+        mkdir build-arm64_32 && cd build-arm64_32
+        cmake ${{ env.COMMON_CMAKE_OPTIONS }} -DPLATFORM=WATCHOS -DARCHS="arm64_32" ..
+        cmake --build . -j 4
+
+    - name: simulator-i386
+      run: |
+        mkdir build-simulator-i386 && cd build-simulator-i386
+        cmake ${{ env.COMMON_CMAKE_OPTIONS }} -DPLATFORM=SIMULATOR_WATCHOS -DARCHS="i386" ..
+        cmake --build . -j 4
+    - name: simulator-x86_64
+      run: |
+        mkdir build-simulator-x86_64 && cd build-simulator-x86_64
+        cmake ${{ env.COMMON_CMAKE_OPTIONS }} -DPLATFORM=SIMULATOR_WATCHOS -DARCHS="x86_64" ..
+        cmake --build . -j 4
+    - name: simulator-arm64
+      run: |
+        mkdir build-simulator-arm64 && cd build-simulator-arm64
+        cmake ${{ env.COMMON_CMAKE_OPTIONS }} -DPLATFORM=SIMULATOR_WATCHOS -DARCHS="arm64" ..
+        cmake --build . -j 4

From d018a159d74f73f6b4d6cd9b01cd3d0fb9e94646 Mon Sep 17 00:00:00 2001
From: nihuini <nihuini@tencent.com>
Date: Mon, 1 Apr 2024 17:42:21 +0800
Subject: [PATCH 03/19] wip

---
 .github/workflows/visionos-cpu.yml |  20 +++---
 .github/workflows/watchos.yml      | 107 ++++++++++++++++++++++++++---
 2 files changed, 106 insertions(+), 21 deletions(-)

diff --git a/.github/workflows/visionos-cpu.yml b/.github/workflows/visionos-cpu.yml
index ac82a4789ccc..0a3cfdda7b12 100644
--- a/.github/workflows/visionos-cpu.yml
+++ b/.github/workflows/visionos-cpu.yml
@@ -10,16 +10,16 @@ on:
     - 'src/*'
     - 'src/layer/*'
     - 'src/layer/arm/**'
-  pull_request:
-    branches: [master]
-    paths:
-    - '.github/workflows/visionos-cpu.yml'
-    - 'toolchains/ios.toolchain.cmake'
-    - 'CMakeLists.txt'
-    - 'cmake/**'
-    - 'src/*'
-    - 'src/layer/*'
-    - 'src/layer/arm/**'
+  # pull_request:
+  #   branches: [master]
+  #   paths:
+  #   - '.github/workflows/visionos-cpu.yml'
+  #   - 'toolchains/ios.toolchain.cmake'
+  #   - 'CMakeLists.txt'
+  #   - 'cmake/**'
+  #   - 'src/*'
+  #   - 'src/layer/*'
+  #   - 'src/layer/arm/**'
 concurrency:
   group: visionos-cpu-${{ github.ref }}
   cancel-in-progress: true
diff --git a/.github/workflows/watchos.yml b/.github/workflows/watchos.yml
index 2ff987891249..adfbb079d31f 100644
--- a/.github/workflows/watchos.yml
+++ b/.github/workflows/watchos.yml
@@ -38,42 +38,127 @@ jobs:
   build:
     runs-on: macos-13
     env:
-      COMMON_CMAKE_OPTIONS: |
+      OPENMP_CMAKE_OPTIONS: |
+        -DCMAKE_TOOLCHAIN_FILE=../../toolchains/ios.toolchain.cmake \
+        -DDEPLOYMENT_TARGET=$WATCHOS_DEPLOYMENT_TARGET \
+        -DENABLE_BITCODE=$ENABLE_BITCODE \
+        -DENABLE_ARC=$ENABLE_ARC \
+        -DENABLE_VISIBILITY=$ENABLE_VISIBILITY \
+        -DCMAKE_INSTALL_PREFIX=install \
+        -DCMAKE_BUILD_TYPE=Release \
+        -DPERL_EXECUTABLE=/usr/local/bin/perl \
+        -DLIBOMP_ENABLE_SHARED=OFF \
+        -DLIBOMP_OMPT_SUPPORT=OFF \
+        -DLIBOMP_USE_HWLOC=OFF \
+
+      NCNN_CMAKE_OPTIONS: |
         -DCMAKE_TOOLCHAIN_FILE=../toolchains/ios.toolchain.cmake \
         -DDEPLOYMENT_TARGET=$WATCHOS_DEPLOYMENT_TARGET \
         -DENABLE_BITCODE=$ENABLE_BITCODE \
         -DENABLE_ARC=$ENABLE_ARC \
         -DENABLE_VISIBILITY=$ENABLE_VISIBILITY \
-        -DNCNN_OPENMP=OFF -DNCNN_THREADS=OFF \
         -DCMAKE_INSTALL_PREFIX=install \
         -DCMAKE_BUILD_TYPE=Release \
+        -DOpenMP_C_FLAGS="-Xclang -fopenmp" -DOpenMP_CXX_FLAGS="-Xclang -fopenmp" \
+        -DOpenMP_C_LIB_NAMES="libomp" -DOpenMP_CXX_LIB_NAMES="libomp" \
+        -DOpenMP_libomp_LIBRARY="libomp.a" \
 
     steps:
     - uses: actions/checkout@v4
 
+    - name: cache-openmp
+      id: cache-openmp
+      uses: actions/cache@v4
+      with:
+        path: openmp-install
+        key: openmp-watchos-install-20240401
+    - name: openmp
+      if: steps.cache-openmp.outputs.cache-hit != 'true'
+      run: |
+        wget https://github.com/llvm/llvm-project/releases/download/llvmorg-18.1.2/openmp-18.1.2.src.tar.xz
+        tar -xf openmp-18.1.2.src.tar.xz
+        cd openmp-18.1.2.src
+        sed -i'' -e '/.size __kmp_unnamed_critical_addr/d' runtime/src/z_Linux_asm.S
+        sed -i'' -e 's/__kmp_unnamed_critical_addr/___kmp_unnamed_critical_addr/g' runtime/src/z_Linux_asm.S
+    - name: openmp-armv7k
+      if: steps.cache-openmp.outputs.cache-hit != 'true'
+      run: |
+        cd openmp-18.1.2.src
+        mkdir -p build-armv7k && cd build-armv7k
+        cmake ${{ env.OPENMP_CMAKE_OPTIONS }} -DPLATFORM=WATCHOS -DARCHS="armv7k" ..
+        cmake --build . -j 4
+        cmake --build . --target install
+    - name: openmp-arm64_32
+      if: steps.cache-openmp.outputs.cache-hit != 'true'
+      run: |
+        cd openmp-18.1.2.src
+        mkdir -p build-arm64_32 && cd build-arm64_32
+        cmake ${{ env.OPENMP_CMAKE_OPTIONS }} -DPLATFORM=WATCHOS -DARCHS="arm64_32" ..
+        cmake --build . -j 4
+        cmake --build . --target install
+    - name: openmp-simulator-x86_64
+      if: steps.cache-openmp.outputs.cache-hit != 'true'
+      run: |
+        cd openmp-18.1.2.src
+        mkdir -p build-simulator-x86_64 && cd build-simulator-x86_64
+        cmake ${{ env.OPENMP_CMAKE_OPTIONS }} -DPLATFORM=SIMULATOR_WATCHOS -DARCHS="x86_64" ..
+        cmake --build . -j 4
+        cmake --build . --target install
+    - name: openmp-simulator-arm64
+      if: steps.cache-openmp.outputs.cache-hit != 'true'
+      run: |
+        cd openmp-18.1.2.src
+        mkdir -p build-simulator-arm64 && cd build-simulator-arm64
+        cmake ${{ env.OPENMP_CMAKE_OPTIONS }} -DPLATFORM=SIMULATOR_WATCHOS -DARCHS="arm64" ..
+        cmake --build . -j 4
+        cmake --build . --target install
+    - name: openmp-merge-fat-library
+      if: steps.cache-openmp.outputs.cache-hit != 'true'
+      run: |
+        mkdir -p $GITHUB_WORKSPACE/openmp-install
+        mkdir -p $GITHUB_WORKSPACE/openmp-install/watchos
+        mkdir -p $GITHUB_WORKSPACE/openmp-install/watchos-simulator
+
+        cp -a openmp-18.1.2.src/build-arm64_32/install/include $GITHUB_WORKSPACE/openmp-install/watchos
+        mkdir -p $GITHUB_WORKSPACE/openmp-install/watchos/lib
+        lipo -create \
+            openmp-18.1.2.src/build-armv7k/install/lib/libomp.a \
+            openmp-18.1.2.src/build-arm64_32/install/lib/libomp.a \
+            -o $GITHUB_WORKSPACE/openmp-install/watchos/lib/libomp.a
+
+        cp -a openmp-18.1.2.src/build-simulator-x86_64/install/include $GITHUB_WORKSPACE/openmp-install/watchos-simulator
+        mkdir -p $GITHUB_WORKSPACE/openmp-install/watchos-simulator/lib
+        lipo -create \
+            openmp-18.1.2.src/build-simulator-x86_64/install/lib/libomp.a \
+            openmp-18.1.2.src/build-simulator-arm64/install/lib/libomp.a \
+            -o $GITHUB_WORKSPACE/openmp-install/watchos-simulator/lib/libomp.a
+
+    - name: install-openmp
+      run: |
+        sudo cp $GITHUB_WORKSPACE/openmp-install/watchos/include/* $DEVELOPER_DIR/Platforms/WatchOS.platform/Developer/SDKs/WatchOS.sdk/usr/include
+        sudo cp $GITHUB_WORKSPACE/openmp-install/watchos/lib/libomp.a $DEVELOPER_DIR/Platforms/WatchOS.platform/Developer/SDKs/WatchOS.sdk/usr/lib
+
+        sudo cp $GITHUB_WORKSPACE/openmp-install/watchos-simulator/include/* $DEVELOPER_DIR/Platforms/WatchSimulator.platform/Developer/SDKs/WatchSimulator.sdk/usr/include
+        sudo cp $GITHUB_WORKSPACE/openmp-install/watchos-simulator/lib/libomp.a $DEVELOPER_DIR/Platforms/WatchSimulator.platform/Developer/SDKs/WatchSimulator.sdk/usr/lib
+
     - name: armv7k
       run: |
         mkdir build-armv7k && cd build-armv7k
-        cmake ${{ env.COMMON_CMAKE_OPTIONS }} -DPLATFORM=WATCHOS -DARCHS="armv7k" ..
+        cmake ${{ env.NCNN_CMAKE_OPTIONS }} -DPLATFORM=WATCHOS -DARCHS="armv7k" ..
         cmake --build . -j 4
     - name: arm64_32
       run: |
         mkdir build-arm64_32 && cd build-arm64_32
-        cmake ${{ env.COMMON_CMAKE_OPTIONS }} -DPLATFORM=WATCHOS -DARCHS="arm64_32" ..
+        cmake ${{ env.NCNN_CMAKE_OPTIONS }} -DPLATFORM=WATCHOS -DARCHS="arm64_32" ..
         cmake --build . -j 4
 
-    - name: simulator-i386
-      run: |
-        mkdir build-simulator-i386 && cd build-simulator-i386
-        cmake ${{ env.COMMON_CMAKE_OPTIONS }} -DPLATFORM=SIMULATOR_WATCHOS -DARCHS="i386" ..
-        cmake --build . -j 4
     - name: simulator-x86_64
       run: |
         mkdir build-simulator-x86_64 && cd build-simulator-x86_64
-        cmake ${{ env.COMMON_CMAKE_OPTIONS }} -DPLATFORM=SIMULATOR_WATCHOS -DARCHS="x86_64" ..
+        cmake ${{ env.NCNN_CMAKE_OPTIONS }} -DPLATFORM=SIMULATOR_WATCHOS -DARCHS="x86_64" ..
         cmake --build . -j 4
     - name: simulator-arm64
       run: |
         mkdir build-simulator-arm64 && cd build-simulator-arm64
-        cmake ${{ env.COMMON_CMAKE_OPTIONS }} -DPLATFORM=SIMULATOR_WATCHOS -DARCHS="arm64" ..
+        cmake ${{ env.NCNN_CMAKE_OPTIONS }} -DPLATFORM=SIMULATOR_WATCHOS -DARCHS="arm64" ..
         cmake --build . -j 4

From 512a4d691315f6ceeed01aa2c189e3fa81744d27 Mon Sep 17 00:00:00 2001
From: nihuini <nihuini@tencent.com>
Date: Mon, 1 Apr 2024 17:46:59 +0800
Subject: [PATCH 04/19] wip

---
 .github/workflows/watchos.yml | 27 ++++++++++++++-------------
 1 file changed, 14 insertions(+), 13 deletions(-)

diff --git a/.github/workflows/watchos.yml b/.github/workflows/watchos.yml
index adfbb079d31f..dd6d5e88d4b2 100644
--- a/.github/workflows/watchos.yml
+++ b/.github/workflows/watchos.yml
@@ -38,6 +38,7 @@ jobs:
   build:
     runs-on: macos-13
     env:
+      OPENMP_VERSION: '16.0.6'
       OPENMP_CMAKE_OPTIONS: |
         -DCMAKE_TOOLCHAIN_FILE=../../toolchains/ios.toolchain.cmake \
         -DDEPLOYMENT_TARGET=$WATCHOS_DEPLOYMENT_TARGET \
@@ -75,15 +76,15 @@ jobs:
     - name: openmp
       if: steps.cache-openmp.outputs.cache-hit != 'true'
       run: |
-        wget https://github.com/llvm/llvm-project/releases/download/llvmorg-18.1.2/openmp-18.1.2.src.tar.xz
-        tar -xf openmp-18.1.2.src.tar.xz
-        cd openmp-18.1.2.src
+        wget https://github.com/llvm/llvm-project/releases/download/llvmorg-${{ env.OPENMP_VERSION }}/openmp-${{ env.OPENMP_VERSION }}.src.tar.xz
+        tar -xf openmp-${{ env.OPENMP_VERSION }}.src.tar.xz
+        cd openmp-${{ env.OPENMP_VERSION }}.src
         sed -i'' -e '/.size __kmp_unnamed_critical_addr/d' runtime/src/z_Linux_asm.S
         sed -i'' -e 's/__kmp_unnamed_critical_addr/___kmp_unnamed_critical_addr/g' runtime/src/z_Linux_asm.S
     - name: openmp-armv7k
       if: steps.cache-openmp.outputs.cache-hit != 'true'
       run: |
-        cd openmp-18.1.2.src
+        cd openmp-${{ env.OPENMP_VERSION }}.src
         mkdir -p build-armv7k && cd build-armv7k
         cmake ${{ env.OPENMP_CMAKE_OPTIONS }} -DPLATFORM=WATCHOS -DARCHS="armv7k" ..
         cmake --build . -j 4
@@ -91,7 +92,7 @@ jobs:
     - name: openmp-arm64_32
       if: steps.cache-openmp.outputs.cache-hit != 'true'
       run: |
-        cd openmp-18.1.2.src
+        cd openmp-${{ env.OPENMP_VERSION }}.src
         mkdir -p build-arm64_32 && cd build-arm64_32
         cmake ${{ env.OPENMP_CMAKE_OPTIONS }} -DPLATFORM=WATCHOS -DARCHS="arm64_32" ..
         cmake --build . -j 4
@@ -99,7 +100,7 @@ jobs:
     - name: openmp-simulator-x86_64
       if: steps.cache-openmp.outputs.cache-hit != 'true'
       run: |
-        cd openmp-18.1.2.src
+        cd openmp-${{ env.OPENMP_VERSION }}.src
         mkdir -p build-simulator-x86_64 && cd build-simulator-x86_64
         cmake ${{ env.OPENMP_CMAKE_OPTIONS }} -DPLATFORM=SIMULATOR_WATCHOS -DARCHS="x86_64" ..
         cmake --build . -j 4
@@ -107,7 +108,7 @@ jobs:
     - name: openmp-simulator-arm64
       if: steps.cache-openmp.outputs.cache-hit != 'true'
       run: |
-        cd openmp-18.1.2.src
+        cd openmp-${{ env.OPENMP_VERSION }}.src
         mkdir -p build-simulator-arm64 && cd build-simulator-arm64
         cmake ${{ env.OPENMP_CMAKE_OPTIONS }} -DPLATFORM=SIMULATOR_WATCHOS -DARCHS="arm64" ..
         cmake --build . -j 4
@@ -119,18 +120,18 @@ jobs:
         mkdir -p $GITHUB_WORKSPACE/openmp-install/watchos
         mkdir -p $GITHUB_WORKSPACE/openmp-install/watchos-simulator
 
-        cp -a openmp-18.1.2.src/build-arm64_32/install/include $GITHUB_WORKSPACE/openmp-install/watchos
+        cp -a openmp-${{ env.OPENMP_VERSION }}.src/build-arm64_32/install/include $GITHUB_WORKSPACE/openmp-install/watchos
         mkdir -p $GITHUB_WORKSPACE/openmp-install/watchos/lib
         lipo -create \
-            openmp-18.1.2.src/build-armv7k/install/lib/libomp.a \
-            openmp-18.1.2.src/build-arm64_32/install/lib/libomp.a \
+            openmp-${{ env.OPENMP_VERSION }}.src/build-armv7k/install/lib/libomp.a \
+            openmp-${{ env.OPENMP_VERSION }}.src/build-arm64_32/install/lib/libomp.a \
             -o $GITHUB_WORKSPACE/openmp-install/watchos/lib/libomp.a
 
-        cp -a openmp-18.1.2.src/build-simulator-x86_64/install/include $GITHUB_WORKSPACE/openmp-install/watchos-simulator
+        cp -a openmp-${{ env.OPENMP_VERSION }}.src/build-simulator-x86_64/install/include $GITHUB_WORKSPACE/openmp-install/watchos-simulator
         mkdir -p $GITHUB_WORKSPACE/openmp-install/watchos-simulator/lib
         lipo -create \
-            openmp-18.1.2.src/build-simulator-x86_64/install/lib/libomp.a \
-            openmp-18.1.2.src/build-simulator-arm64/install/lib/libomp.a \
+            openmp-${{ env.OPENMP_VERSION }}.src/build-simulator-x86_64/install/lib/libomp.a \
+            openmp-${{ env.OPENMP_VERSION }}.src/build-simulator-arm64/install/lib/libomp.a \
             -o $GITHUB_WORKSPACE/openmp-install/watchos-simulator/lib/libomp.a
 
     - name: install-openmp

From c185c01838552fdfcdf1cddd91c321b5a7e0de11 Mon Sep 17 00:00:00 2001
From: nihuini <nihuini@tencent.com>
Date: Mon, 1 Apr 2024 17:50:00 +0800
Subject: [PATCH 05/19] wip

---
 .github/workflows/watchos.yml | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/.github/workflows/watchos.yml b/.github/workflows/watchos.yml
index dd6d5e88d4b2..ee30d2ed3b32 100644
--- a/.github/workflows/watchos.yml
+++ b/.github/workflows/watchos.yml
@@ -38,7 +38,7 @@ jobs:
   build:
     runs-on: macos-13
     env:
-      OPENMP_VERSION: '16.0.6'
+      OPENMP_VERSION: '11.0.0'
       OPENMP_CMAKE_OPTIONS: |
         -DCMAKE_TOOLCHAIN_FILE=../../toolchains/ios.toolchain.cmake \
         -DDEPLOYMENT_TARGET=$WATCHOS_DEPLOYMENT_TARGET \

From 9093ff31dd0ea5d6280827eb7dba38cd1b678972 Mon Sep 17 00:00:00 2001
From: nihuini <nihuini@tencent.com>
Date: Mon, 1 Apr 2024 19:08:58 +0800
Subject: [PATCH 06/19] wip

---
 .github/workflows/visionos-cpu.yml |  73 --------------
 .github/workflows/visionos.yml     | 148 +++++++++++++++++++++++++++++
 .github/workflows/watchos.yml      |  92 +-----------------
 3 files changed, 149 insertions(+), 164 deletions(-)
 delete mode 100644 .github/workflows/visionos-cpu.yml
 create mode 100644 .github/workflows/visionos.yml

diff --git a/.github/workflows/visionos-cpu.yml b/.github/workflows/visionos-cpu.yml
deleted file mode 100644
index 0a3cfdda7b12..000000000000
--- a/.github/workflows/visionos-cpu.yml
+++ /dev/null
@@ -1,73 +0,0 @@
-name: visionos-cpu
-on:
-  push:
-    branches: [master]
-    paths:
-    - '.github/workflows/visionos-cpu.yml'
-    - 'toolchains/ios.toolchain.cmake'
-    - 'CMakeLists.txt'
-    - 'cmake/**'
-    - 'src/*'
-    - 'src/layer/*'
-    - 'src/layer/arm/**'
-  # pull_request:
-  #   branches: [master]
-  #   paths:
-  #   - '.github/workflows/visionos-cpu.yml'
-  #   - 'toolchains/ios.toolchain.cmake'
-  #   - 'CMakeLists.txt'
-  #   - 'cmake/**'
-  #   - 'src/*'
-  #   - 'src/layer/*'
-  #   - 'src/layer/arm/**'
-concurrency:
-  group: visionos-cpu-${{ github.ref }}
-  cancel-in-progress: true
-env:
-  DEVELOPER_DIR: /Applications/Xcode_15.2.app/Contents/Developer
-  VISIONOS_DEPLOYMENT_TARGET: '1.0'
-  ENABLE_BITCODE: OFF
-  ENABLE_ARC: OFF
-  ENABLE_VISIBILITY: OFF
-permissions:
-  contents: read
-
-jobs:
-  visionos:
-    runs-on: macos-13
-    steps:
-    - uses: actions/checkout@v4
-
-    - name: build-arm64
-      run: |
-        mkdir build-arm64 && cd build-arm64
-        cmake -DCMAKE_TOOLCHAIN_FILE=../toolchains/ios.toolchain.cmake -DPLATFORM=VISIONOS -DARCHS="arm64" \
-            -DDEPLOYMENT_TARGET=$VISIONOS_DEPLOYMENT_TARGET -DENABLE_BITCODE=$ENABLE_BITCODE -DENABLE_ARC=$ENABLE_ARC -DENABLE_VISIBILITY=$ENABLE_VISIBILITY \
-            -DNCNN_OPENMP=OFF -DNCNN_THREADS=OFF \
-            -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=install \
-            ..
-        cmake --build . -j 3
-
-  visionos-simulator:
-    runs-on: macos-13
-    steps:
-    - uses: actions/checkout@v4
-
-    - name: build-x86_64
-      run: |
-        mkdir build-x86_64 && cd build-x86_64
-        cmake -DCMAKE_TOOLCHAIN_FILE=../toolchains/ios.toolchain.cmake -DPLATFORM=SIMULATOR_VISIONOS -DARCHS="x86_64" \
-            -DDEPLOYMENT_TARGET=$VISIONOS_DEPLOYMENT_TARGET -DENABLE_BITCODE=$ENABLE_BITCODE -DENABLE_ARC=$ENABLE_ARC -DENABLE_VISIBILITY=$ENABLE_VISIBILITY \
-            -DNCNN_OPENMP=OFF -DNCNN_THREADS=OFF \
-            -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=install \
-            ..
-        cmake --build . -j 3
-    - name: build-arm64
-      run: |
-        mkdir build-arm64 && cd build-arm64
-        cmake -DCMAKE_TOOLCHAIN_FILE=../toolchains/ios.toolchain.cmake -DPLATFORM=SIMULATOR_VISIONOS -DARCHS="arm64" \
-            -DDEPLOYMENT_TARGET=$VISIONOS_DEPLOYMENT_TARGET -DENABLE_BITCODE=$ENABLE_BITCODE -DENABLE_ARC=$ENABLE_ARC -DENABLE_VISIBILITY=$ENABLE_VISIBILITY \
-            -DNCNN_OPENMP=OFF -DNCNN_THREADS=OFF \
-            -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=install \
-            ..
-        cmake --build . -j 3
diff --git a/.github/workflows/visionos.yml b/.github/workflows/visionos.yml
new file mode 100644
index 000000000000..3fe1df876264
--- /dev/null
+++ b/.github/workflows/visionos.yml
@@ -0,0 +1,148 @@
+name: visionos
+on:
+  push:
+    branches: [master]
+    paths:
+    - '.github/workflows/visionos.yml'
+    - 'toolchains/ios.toolchain.cmake'
+    - 'CMakeLists.txt'
+    - 'cmake/**'
+    - 'src/*'
+    - 'src/layer/*'
+    - 'src/layer/arm/**'
+    - 'src/layer/x86/**'
+  pull_request:
+    branches: [master]
+    paths:
+    - '.github/workflows/visionos.yml'
+    - 'toolchains/ios.toolchain.cmake'
+    - 'CMakeLists.txt'
+    - 'cmake/**'
+    - 'src/*'
+    - 'src/layer/*'
+    - 'src/layer/arm/**'
+    - 'src/layer/x86/**'
+concurrency:
+  group: visionos-${{ github.ref }}
+  cancel-in-progress: true
+env:
+  DEVELOPER_DIR: /Applications/Xcode_15.2.app/Contents/Developer
+  VISIONOS_DEPLOYMENT_TARGET: '1.0'
+  ENABLE_BITCODE: OFF
+  ENABLE_ARC: OFF
+  ENABLE_VISIBILITY: OFF
+permissions:
+  contents: read
+
+jobs:
+  build:
+    runs-on: macos-13
+    env:
+      OPENMP_VERSION: '11.0.1'
+      OPENMP_CMAKE_OPTIONS: |
+        -DCMAKE_TOOLCHAIN_FILE=../../toolchains/ios.toolchain.cmake \
+        -DDEPLOYMENT_TARGET=$VISIONOS_DEPLOYMENT_TARGET \
+        -DENABLE_BITCODE=$ENABLE_BITCODE \
+        -DENABLE_ARC=$ENABLE_ARC \
+        -DENABLE_VISIBILITY=$ENABLE_VISIBILITY \
+        -DCMAKE_INSTALL_PREFIX=install \
+        -DCMAKE_BUILD_TYPE=Release \
+        -DPERL_EXECUTABLE=/usr/local/bin/perl \
+        -DLIBOMP_ENABLE_SHARED=OFF \
+        -DLIBOMP_OMPT_SUPPORT=OFF \
+        -DLIBOMP_USE_HWLOC=OFF \
+
+      NCNN_CMAKE_OPTIONS: |
+        -DCMAKE_TOOLCHAIN_FILE=../toolchains/ios.toolchain.cmake \
+        -DDEPLOYMENT_TARGET=$VISIONOS_DEPLOYMENT_TARGET \
+        -DENABLE_BITCODE=$ENABLE_BITCODE \
+        -DENABLE_ARC=$ENABLE_ARC \
+        -DENABLE_VISIBILITY=$ENABLE_VISIBILITY \
+        -DCMAKE_INSTALL_PREFIX=install \
+        -DCMAKE_BUILD_TYPE=Release \
+        -DOpenMP_C_FLAGS="-Xclang -fopenmp" -DOpenMP_CXX_FLAGS="-Xclang -fopenmp" \
+        -DOpenMP_C_LIB_NAMES="libomp" -DOpenMP_CXX_LIB_NAMES="libomp" \
+        -DOpenMP_libomp_LIBRARY="libomp.a" \
+
+    steps:
+    - uses: actions/checkout@v4
+
+    - name: cache-openmp
+      id: cache-openmp
+      uses: actions/cache@v4
+      with:
+        path: openmp-install
+        key: openmp-visionos-install-20240401
+    - name: openmp
+      if: steps.cache-openmp.outputs.cache-hit != 'true'
+      run: |
+        wget https://github.com/llvm/llvm-project/releases/download/llvmorg-${{ env.OPENMP_VERSION }}/openmp-${{ env.OPENMP_VERSION }}.src.tar.xz
+        tar -xf openmp-${{ env.OPENMP_VERSION }}.src.tar.xz
+        cd openmp-${{ env.OPENMP_VERSION }}.src
+        sed -i'' -e '/.size __kmp_unnamed_critical_addr/d' runtime/src/z_Linux_asm.S
+        sed -i'' -e 's/__kmp_unnamed_critical_addr/___kmp_unnamed_critical_addr/g' runtime/src/z_Linux_asm.S
+    - name: openmp-arm64
+      if: steps.cache-openmp.outputs.cache-hit != 'true'
+      run: |
+        cd openmp-${{ env.OPENMP_VERSION }}.src
+        mkdir -p build-arm64 && cd build-arm64
+        cmake ${{ env.OPENMP_CMAKE_OPTIONS }} -DPLATFORM=WATCHOS -DARCHS="arm64" ..
+        cmake --build . -j 4
+        cmake --build . --target install
+    - name: openmp-simulator-x86_64
+      if: steps.cache-openmp.outputs.cache-hit != 'true'
+      run: |
+        cd openmp-${{ env.OPENMP_VERSION }}.src
+        mkdir -p build-simulator-x86_64 && cd build-simulator-x86_64
+        cmake ${{ env.OPENMP_CMAKE_OPTIONS }} -DPLATFORM=SIMULATOR_WATCHOS -DARCHS="x86_64" ..
+        cmake --build . -j 4
+        cmake --build . --target install
+    - name: openmp-simulator-arm64
+      if: steps.cache-openmp.outputs.cache-hit != 'true'
+      run: |
+        cd openmp-${{ env.OPENMP_VERSION }}.src
+        mkdir -p build-simulator-arm64 && cd build-simulator-arm64
+        cmake ${{ env.OPENMP_CMAKE_OPTIONS }} -DPLATFORM=SIMULATOR_WATCHOS -DARCHS="arm64" ..
+        cmake --build . -j 4
+        cmake --build . --target install
+    - name: openmp-merge-fat-library
+      if: steps.cache-openmp.outputs.cache-hit != 'true'
+      run: |
+        mkdir -p $GITHUB_WORKSPACE/openmp-install
+        mkdir -p $GITHUB_WORKSPACE/openmp-install/visionos
+        mkdir -p $GITHUB_WORKSPACE/openmp-install/visionos-simulator
+
+        cp -a openmp-${{ env.OPENMP_VERSION }}.src/build-arm64/install/include $GITHUB_WORKSPACE/openmp-install/visionos
+        mkdir -p $GITHUB_WORKSPACE/openmp-install/visionos/lib
+        cp openmp-${{ env.OPENMP_VERSION }}.src/build-arm64/install/lib/libomp.a $GITHUB_WORKSPACE/openmp-install/visionos/lib/libomp.a
+
+        cp -a openmp-${{ env.OPENMP_VERSION }}.src/build-simulator-x86_64/install/include $GITHUB_WORKSPACE/openmp-install/visionos-simulator
+        mkdir -p $GITHUB_WORKSPACE/openmp-install/visionos-simulator/lib
+        lipo -create \
+            openmp-${{ env.OPENMP_VERSION }}.src/build-simulator-x86_64/install/lib/libomp.a \
+            openmp-${{ env.OPENMP_VERSION }}.src/build-simulator-arm64/install/lib/libomp.a \
+            -o $GITHUB_WORKSPACE/openmp-install/visionos-simulator/lib/libomp.a
+
+    - name: install-openmp
+      run: |
+        sudo cp $GITHUB_WORKSPACE/openmp-install/visionos/include/* $DEVELOPER_DIR/Platforms/XROS.platform/Developer/SDKs/XROS.sdk/usr/include
+        sudo cp $GITHUB_WORKSPACE/openmp-install/visionos/lib/libomp.a $DEVELOPER_DIR/Platforms/XROS.platform/Developer/SDKs/XROS.sdk/usr/lib
+
+        sudo cp $GITHUB_WORKSPACE/openmp-install/visionos-simulator/include/* $DEVELOPER_DIR/Platforms/XRSimulator.platform/Developer/SDKs/XRSimulator.sdk/usr/include
+        sudo cp $GITHUB_WORKSPACE/openmp-install/visionos-simulator/lib/libomp.a $DEVELOPER_DIR/Platforms/XRSimulator.platform/Developer/SDKs/XRSimulator.sdk/usr/lib
+
+    - name: arm64
+      run: |
+        mkdir build-arm64 && cd build-arm64
+        cmake ${{ env.NCNN_CMAKE_OPTIONS }} -DPLATFORM=VISIONOS -DARCHS="arm64" ..
+        cmake --build . -j 4
+    - name: simulator-x86_64
+      run: |
+        mkdir build-simulator-x86_64 && cd build-simulator-x86_64
+        cmake ${{ env.NCNN_CMAKE_OPTIONS }} -DPLATFORM=SIMULATOR_VISIONOS -DARCHS="x86_64" ..
+        cmake --build . -j 4
+    - name: simulator-arm64
+      run: |
+        mkdir build-simulator-arm64 && cd build-simulator-arm64
+        cmake ${{ env.NCNN_CMAKE_OPTIONS }} -DPLATFORM=SIMULATOR_VISIONOS -DARCHS="arm64" ..
+        cmake --build . -j 4
diff --git a/.github/workflows/watchos.yml b/.github/workflows/watchos.yml
index ee30d2ed3b32..b809838b4ec1 100644
--- a/.github/workflows/watchos.yml
+++ b/.github/workflows/watchos.yml
@@ -26,7 +26,7 @@ concurrency:
   group: watchos-${{ github.ref }}
   cancel-in-progress: true
 env:
-  DEVELOPER_DIR: /Applications/Xcode_14.3.1.app/Contents/Developer
+  DEVELOPER_DIR: /Applications/Xcode_15.2.app/Contents/Developer
   WATCHOS_DEPLOYMENT_TARGET: '2.0'
   ENABLE_BITCODE: OFF
   ENABLE_ARC: OFF
@@ -38,20 +38,6 @@ jobs:
   build:
     runs-on: macos-13
     env:
-      OPENMP_VERSION: '11.0.0'
-      OPENMP_CMAKE_OPTIONS: |
-        -DCMAKE_TOOLCHAIN_FILE=../../toolchains/ios.toolchain.cmake \
-        -DDEPLOYMENT_TARGET=$WATCHOS_DEPLOYMENT_TARGET \
-        -DENABLE_BITCODE=$ENABLE_BITCODE \
-        -DENABLE_ARC=$ENABLE_ARC \
-        -DENABLE_VISIBILITY=$ENABLE_VISIBILITY \
-        -DCMAKE_INSTALL_PREFIX=install \
-        -DCMAKE_BUILD_TYPE=Release \
-        -DPERL_EXECUTABLE=/usr/local/bin/perl \
-        -DLIBOMP_ENABLE_SHARED=OFF \
-        -DLIBOMP_OMPT_SUPPORT=OFF \
-        -DLIBOMP_USE_HWLOC=OFF \
-
       NCNN_CMAKE_OPTIONS: |
         -DCMAKE_TOOLCHAIN_FILE=../toolchains/ios.toolchain.cmake \
         -DDEPLOYMENT_TARGET=$WATCHOS_DEPLOYMENT_TARGET \
@@ -67,81 +53,6 @@ jobs:
     steps:
     - uses: actions/checkout@v4
 
-    - name: cache-openmp
-      id: cache-openmp
-      uses: actions/cache@v4
-      with:
-        path: openmp-install
-        key: openmp-watchos-install-20240401
-    - name: openmp
-      if: steps.cache-openmp.outputs.cache-hit != 'true'
-      run: |
-        wget https://github.com/llvm/llvm-project/releases/download/llvmorg-${{ env.OPENMP_VERSION }}/openmp-${{ env.OPENMP_VERSION }}.src.tar.xz
-        tar -xf openmp-${{ env.OPENMP_VERSION }}.src.tar.xz
-        cd openmp-${{ env.OPENMP_VERSION }}.src
-        sed -i'' -e '/.size __kmp_unnamed_critical_addr/d' runtime/src/z_Linux_asm.S
-        sed -i'' -e 's/__kmp_unnamed_critical_addr/___kmp_unnamed_critical_addr/g' runtime/src/z_Linux_asm.S
-    - name: openmp-armv7k
-      if: steps.cache-openmp.outputs.cache-hit != 'true'
-      run: |
-        cd openmp-${{ env.OPENMP_VERSION }}.src
-        mkdir -p build-armv7k && cd build-armv7k
-        cmake ${{ env.OPENMP_CMAKE_OPTIONS }} -DPLATFORM=WATCHOS -DARCHS="armv7k" ..
-        cmake --build . -j 4
-        cmake --build . --target install
-    - name: openmp-arm64_32
-      if: steps.cache-openmp.outputs.cache-hit != 'true'
-      run: |
-        cd openmp-${{ env.OPENMP_VERSION }}.src
-        mkdir -p build-arm64_32 && cd build-arm64_32
-        cmake ${{ env.OPENMP_CMAKE_OPTIONS }} -DPLATFORM=WATCHOS -DARCHS="arm64_32" ..
-        cmake --build . -j 4
-        cmake --build . --target install
-    - name: openmp-simulator-x86_64
-      if: steps.cache-openmp.outputs.cache-hit != 'true'
-      run: |
-        cd openmp-${{ env.OPENMP_VERSION }}.src
-        mkdir -p build-simulator-x86_64 && cd build-simulator-x86_64
-        cmake ${{ env.OPENMP_CMAKE_OPTIONS }} -DPLATFORM=SIMULATOR_WATCHOS -DARCHS="x86_64" ..
-        cmake --build . -j 4
-        cmake --build . --target install
-    - name: openmp-simulator-arm64
-      if: steps.cache-openmp.outputs.cache-hit != 'true'
-      run: |
-        cd openmp-${{ env.OPENMP_VERSION }}.src
-        mkdir -p build-simulator-arm64 && cd build-simulator-arm64
-        cmake ${{ env.OPENMP_CMAKE_OPTIONS }} -DPLATFORM=SIMULATOR_WATCHOS -DARCHS="arm64" ..
-        cmake --build . -j 4
-        cmake --build . --target install
-    - name: openmp-merge-fat-library
-      if: steps.cache-openmp.outputs.cache-hit != 'true'
-      run: |
-        mkdir -p $GITHUB_WORKSPACE/openmp-install
-        mkdir -p $GITHUB_WORKSPACE/openmp-install/watchos
-        mkdir -p $GITHUB_WORKSPACE/openmp-install/watchos-simulator
-
-        cp -a openmp-${{ env.OPENMP_VERSION }}.src/build-arm64_32/install/include $GITHUB_WORKSPACE/openmp-install/watchos
-        mkdir -p $GITHUB_WORKSPACE/openmp-install/watchos/lib
-        lipo -create \
-            openmp-${{ env.OPENMP_VERSION }}.src/build-armv7k/install/lib/libomp.a \
-            openmp-${{ env.OPENMP_VERSION }}.src/build-arm64_32/install/lib/libomp.a \
-            -o $GITHUB_WORKSPACE/openmp-install/watchos/lib/libomp.a
-
-        cp -a openmp-${{ env.OPENMP_VERSION }}.src/build-simulator-x86_64/install/include $GITHUB_WORKSPACE/openmp-install/watchos-simulator
-        mkdir -p $GITHUB_WORKSPACE/openmp-install/watchos-simulator/lib
-        lipo -create \
-            openmp-${{ env.OPENMP_VERSION }}.src/build-simulator-x86_64/install/lib/libomp.a \
-            openmp-${{ env.OPENMP_VERSION }}.src/build-simulator-arm64/install/lib/libomp.a \
-            -o $GITHUB_WORKSPACE/openmp-install/watchos-simulator/lib/libomp.a
-
-    - name: install-openmp
-      run: |
-        sudo cp $GITHUB_WORKSPACE/openmp-install/watchos/include/* $DEVELOPER_DIR/Platforms/WatchOS.platform/Developer/SDKs/WatchOS.sdk/usr/include
-        sudo cp $GITHUB_WORKSPACE/openmp-install/watchos/lib/libomp.a $DEVELOPER_DIR/Platforms/WatchOS.platform/Developer/SDKs/WatchOS.sdk/usr/lib
-
-        sudo cp $GITHUB_WORKSPACE/openmp-install/watchos-simulator/include/* $DEVELOPER_DIR/Platforms/WatchSimulator.platform/Developer/SDKs/WatchSimulator.sdk/usr/include
-        sudo cp $GITHUB_WORKSPACE/openmp-install/watchos-simulator/lib/libomp.a $DEVELOPER_DIR/Platforms/WatchSimulator.platform/Developer/SDKs/WatchSimulator.sdk/usr/lib
-
     - name: armv7k
       run: |
         mkdir build-armv7k && cd build-armv7k
@@ -152,7 +63,6 @@ jobs:
         mkdir build-arm64_32 && cd build-arm64_32
         cmake ${{ env.NCNN_CMAKE_OPTIONS }} -DPLATFORM=WATCHOS -DARCHS="arm64_32" ..
         cmake --build . -j 4
-
     - name: simulator-x86_64
       run: |
         mkdir build-simulator-x86_64 && cd build-simulator-x86_64

From 6341aed89ef607af90ca2ac69bedc3ca1b0b014b Mon Sep 17 00:00:00 2001
From: nihuini <nihuini@tencent.com>
Date: Mon, 1 Apr 2024 19:11:28 +0800
Subject: [PATCH 07/19] wip

---
 .github/workflows/visionos.yml | 6 +++---
 .github/workflows/watchos.yml  | 4 +---
 2 files changed, 4 insertions(+), 6 deletions(-)

diff --git a/.github/workflows/visionos.yml b/.github/workflows/visionos.yml
index 3fe1df876264..845754754c2c 100644
--- a/.github/workflows/visionos.yml
+++ b/.github/workflows/visionos.yml
@@ -86,7 +86,7 @@ jobs:
       run: |
         cd openmp-${{ env.OPENMP_VERSION }}.src
         mkdir -p build-arm64 && cd build-arm64
-        cmake ${{ env.OPENMP_CMAKE_OPTIONS }} -DPLATFORM=WATCHOS -DARCHS="arm64" ..
+        cmake ${{ env.OPENMP_CMAKE_OPTIONS }} -DPLATFORM=VISIONOS -DARCHS="arm64" ..
         cmake --build . -j 4
         cmake --build . --target install
     - name: openmp-simulator-x86_64
@@ -94,7 +94,7 @@ jobs:
       run: |
         cd openmp-${{ env.OPENMP_VERSION }}.src
         mkdir -p build-simulator-x86_64 && cd build-simulator-x86_64
-        cmake ${{ env.OPENMP_CMAKE_OPTIONS }} -DPLATFORM=SIMULATOR_WATCHOS -DARCHS="x86_64" ..
+        cmake ${{ env.OPENMP_CMAKE_OPTIONS }} -DPLATFORM=SIMULATOR_VISIONOS -DARCHS="x86_64" ..
         cmake --build . -j 4
         cmake --build . --target install
     - name: openmp-simulator-arm64
@@ -102,7 +102,7 @@ jobs:
       run: |
         cd openmp-${{ env.OPENMP_VERSION }}.src
         mkdir -p build-simulator-arm64 && cd build-simulator-arm64
-        cmake ${{ env.OPENMP_CMAKE_OPTIONS }} -DPLATFORM=SIMULATOR_WATCHOS -DARCHS="arm64" ..
+        cmake ${{ env.OPENMP_CMAKE_OPTIONS }} -DPLATFORM=SIMULATOR_VISIONOS -DARCHS="arm64" ..
         cmake --build . -j 4
         cmake --build . --target install
     - name: openmp-merge-fat-library
diff --git a/.github/workflows/watchos.yml b/.github/workflows/watchos.yml
index b809838b4ec1..5da9b0f45084 100644
--- a/.github/workflows/watchos.yml
+++ b/.github/workflows/watchos.yml
@@ -46,9 +46,7 @@ jobs:
         -DENABLE_VISIBILITY=$ENABLE_VISIBILITY \
         -DCMAKE_INSTALL_PREFIX=install \
         -DCMAKE_BUILD_TYPE=Release \
-        -DOpenMP_C_FLAGS="-Xclang -fopenmp" -DOpenMP_CXX_FLAGS="-Xclang -fopenmp" \
-        -DOpenMP_C_LIB_NAMES="libomp" -DOpenMP_CXX_LIB_NAMES="libomp" \
-        -DOpenMP_libomp_LIBRARY="libomp.a" \
+        -DNCNN_OPENMP=OFF -DNCNN_THREADS=OFF \
 
     steps:
     - uses: actions/checkout@v4

From 48ef145dfe610a907b5753bff3bd22a562a49dfa Mon Sep 17 00:00:00 2001
From: nihuini <nihuini@tencent.com>
Date: Mon, 1 Apr 2024 19:16:24 +0800
Subject: [PATCH 08/19] wip

---
 .github/workflows/visionos.yml |  2 +-
 .github/workflows/watchos.yml  | 94 +++++++++++++++++++++++++++++++++-
 2 files changed, 94 insertions(+), 2 deletions(-)

diff --git a/.github/workflows/visionos.yml b/.github/workflows/visionos.yml
index 845754754c2c..90f4c8092870 100644
--- a/.github/workflows/visionos.yml
+++ b/.github/workflows/visionos.yml
@@ -38,7 +38,7 @@ jobs:
   build:
     runs-on: macos-13
     env:
-      OPENMP_VERSION: '11.0.1'
+      OPENMP_VERSION: '14.0.6'
       OPENMP_CMAKE_OPTIONS: |
         -DCMAKE_TOOLCHAIN_FILE=../../toolchains/ios.toolchain.cmake \
         -DDEPLOYMENT_TARGET=$VISIONOS_DEPLOYMENT_TARGET \
diff --git a/.github/workflows/watchos.yml b/.github/workflows/watchos.yml
index 5da9b0f45084..4f0802852bb0 100644
--- a/.github/workflows/watchos.yml
+++ b/.github/workflows/watchos.yml
@@ -38,6 +38,20 @@ jobs:
   build:
     runs-on: macos-13
     env:
+      OPENMP_VERSION: '14.0.6'
+      OPENMP_CMAKE_OPTIONS: |
+        -DCMAKE_TOOLCHAIN_FILE=../../toolchains/ios.toolchain.cmake \
+        -DDEPLOYMENT_TARGET=$WATCHOS_DEPLOYMENT_TARGET \
+        -DENABLE_BITCODE=$ENABLE_BITCODE \
+        -DENABLE_ARC=$ENABLE_ARC \
+        -DENABLE_VISIBILITY=$ENABLE_VISIBILITY \
+        -DCMAKE_INSTALL_PREFIX=install \
+        -DCMAKE_BUILD_TYPE=Release \
+        -DPERL_EXECUTABLE=/usr/local/bin/perl \
+        -DLIBOMP_ENABLE_SHARED=OFF \
+        -DLIBOMP_OMPT_SUPPORT=OFF \
+        -DLIBOMP_USE_HWLOC=OFF \
+
       NCNN_CMAKE_OPTIONS: |
         -DCMAKE_TOOLCHAIN_FILE=../toolchains/ios.toolchain.cmake \
         -DDEPLOYMENT_TARGET=$WATCHOS_DEPLOYMENT_TARGET \
@@ -46,11 +60,88 @@ jobs:
         -DENABLE_VISIBILITY=$ENABLE_VISIBILITY \
         -DCMAKE_INSTALL_PREFIX=install \
         -DCMAKE_BUILD_TYPE=Release \
-        -DNCNN_OPENMP=OFF -DNCNN_THREADS=OFF \
+        -DOpenMP_C_FLAGS="-Xclang -fopenmp" -DOpenMP_CXX_FLAGS="-Xclang -fopenmp" \
+        -DOpenMP_C_LIB_NAMES="libomp" -DOpenMP_CXX_LIB_NAMES="libomp" \
+        -DOpenMP_libomp_LIBRARY="libomp.a" \
 
     steps:
     - uses: actions/checkout@v4
 
+    - name: cache-openmp
+      id: cache-openmp
+      uses: actions/cache@v4
+      with:
+        path: openmp-install
+        key: openmp-watchos-install-20240401
+    - name: openmp
+      if: steps.cache-openmp.outputs.cache-hit != 'true'
+      run: |
+        wget https://github.com/llvm/llvm-project/releases/download/llvmorg-${{ env.OPENMP_VERSION }}/openmp-${{ env.OPENMP_VERSION }}.src.tar.xz
+        tar -xf openmp-${{ env.OPENMP_VERSION }}.src.tar.xz
+        cd openmp-${{ env.OPENMP_VERSION }}.src
+        sed -i'' -e '/.size __kmp_unnamed_critical_addr/d' runtime/src/z_Linux_asm.S
+        sed -i'' -e 's/__kmp_unnamed_critical_addr/___kmp_unnamed_critical_addr/g' runtime/src/z_Linux_asm.S
+    - name: openmp-armv7k
+      if: steps.cache-openmp.outputs.cache-hit != 'true'
+      run: |
+        cd openmp-${{ env.OPENMP_VERSION }}.src
+        mkdir -p build-armv7k && cd build-armv7k
+        cmake ${{ env.OPENMP_CMAKE_OPTIONS }} -DPLATFORM=WATCHOS -DARCHS="armv7k" ..
+        cmake --build . -j 4
+        cmake --build . --target install
+    - name: openmp-arm64_32
+      if: steps.cache-openmp.outputs.cache-hit != 'true'
+      run: |
+        cd openmp-${{ env.OPENMP_VERSION }}.src
+        mkdir -p build-arm64_32 && cd build-arm64_32
+        cmake ${{ env.OPENMP_CMAKE_OPTIONS }} -DPLATFORM=WATCHOS -DARCHS="arm64_32" ..
+        cmake --build . -j 4
+        cmake --build . --target install
+    - name: openmp-simulator-x86_64
+      if: steps.cache-openmp.outputs.cache-hit != 'true'
+      run: |
+        cd openmp-${{ env.OPENMP_VERSION }}.src
+        mkdir -p build-simulator-x86_64 && cd build-simulator-x86_64
+        cmake ${{ env.OPENMP_CMAKE_OPTIONS }} -DPLATFORM=SIMULATOR_WATCHOS -DARCHS="x86_64" ..
+        cmake --build . -j 4
+        cmake --build . --target install
+    - name: openmp-simulator-arm64
+      if: steps.cache-openmp.outputs.cache-hit != 'true'
+      run: |
+        cd openmp-${{ env.OPENMP_VERSION }}.src
+        mkdir -p build-simulator-arm64 && cd build-simulator-arm64
+        cmake ${{ env.OPENMP_CMAKE_OPTIONS }} -DPLATFORM=SIMULATOR_WATCHOS -DARCHS="arm64" ..
+        cmake --build . -j 4
+        cmake --build . --target install
+    - name: openmp-merge-fat-library
+      if: steps.cache-openmp.outputs.cache-hit != 'true'
+      run: |
+        mkdir -p $GITHUB_WORKSPACE/openmp-install
+        mkdir -p $GITHUB_WORKSPACE/openmp-install/watchos
+        mkdir -p $GITHUB_WORKSPACE/openmp-install/watchos-simulator
+
+        cp -a openmp-${{ env.OPENMP_VERSION }}.src/build-arm64_32/install/include $GITHUB_WORKSPACE/openmp-install/watchos
+        mkdir -p $GITHUB_WORKSPACE/openmp-install/watchos/lib
+        lipo -create \
+            openmp-${{ env.OPENMP_VERSION }}.src/build-armv7k/install/lib/libomp.a \
+            openmp-${{ env.OPENMP_VERSION }}.src/build-arm64_32/install/lib/libomp.a \
+            -o $GITHUB_WORKSPACE/openmp-install/watchos/lib/libomp.a
+
+        cp -a openmp-${{ env.OPENMP_VERSION }}.src/build-simulator-x86_64/install/include $GITHUB_WORKSPACE/openmp-install/watchos-simulator
+        mkdir -p $GITHUB_WORKSPACE/openmp-install/watchos-simulator/lib
+        lipo -create \
+            openmp-${{ env.OPENMP_VERSION }}.src/build-simulator-x86_64/install/lib/libomp.a \
+            openmp-${{ env.OPENMP_VERSION }}.src/build-simulator-arm64/install/lib/libomp.a \
+            -o $GITHUB_WORKSPACE/openmp-install/watchos-simulator/lib/libomp.a
+
+    - name: install-openmp
+      run: |
+        sudo cp $GITHUB_WORKSPACE/openmp-install/watchos/include/* $DEVELOPER_DIR/Platforms/WatchOS.platform/Developer/SDKs/WatchOS.sdk/usr/include
+        sudo cp $GITHUB_WORKSPACE/openmp-install/watchos/lib/libomp.a $DEVELOPER_DIR/Platforms/WatchOS.platform/Developer/SDKs/WatchOS.sdk/usr/lib
+
+        sudo cp $GITHUB_WORKSPACE/openmp-install/watchos-simulator/include/* $DEVELOPER_DIR/Platforms/WatchSimulator.platform/Developer/SDKs/WatchSimulator.sdk/usr/include
+        sudo cp $GITHUB_WORKSPACE/openmp-install/watchos-simulator/lib/libomp.a $DEVELOPER_DIR/Platforms/WatchSimulator.platform/Developer/SDKs/WatchSimulator.sdk/usr/lib
+
     - name: armv7k
       run: |
         mkdir build-armv7k && cd build-armv7k
@@ -61,6 +152,7 @@ jobs:
         mkdir build-arm64_32 && cd build-arm64_32
         cmake ${{ env.NCNN_CMAKE_OPTIONS }} -DPLATFORM=WATCHOS -DARCHS="arm64_32" ..
         cmake --build . -j 4
+
     - name: simulator-x86_64
       run: |
         mkdir build-simulator-x86_64 && cd build-simulator-x86_64

From 27ff00583baddabb684d21c589570463ebbf0722 Mon Sep 17 00:00:00 2001
From: nihuini <nihuini@tencent.com>
Date: Mon, 1 Apr 2024 19:56:39 +0800
Subject: [PATCH 09/19] wip

---
 .github/workflows/watchos.yml | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/.github/workflows/watchos.yml b/.github/workflows/watchos.yml
index 4f0802852bb0..57d87f7023aa 100644
--- a/.github/workflows/watchos.yml
+++ b/.github/workflows/watchos.yml
@@ -27,7 +27,7 @@ concurrency:
   cancel-in-progress: true
 env:
   DEVELOPER_DIR: /Applications/Xcode_15.2.app/Contents/Developer
-  WATCHOS_DEPLOYMENT_TARGET: '2.0'
+  WATCHOS_DEPLOYMENT_TARGET: '6.0'
   ENABLE_BITCODE: OFF
   ENABLE_ARC: OFF
   ENABLE_VISIBILITY: OFF
@@ -81,6 +81,8 @@ jobs:
         cd openmp-${{ env.OPENMP_VERSION }}.src
         sed -i'' -e '/.size __kmp_unnamed_critical_addr/d' runtime/src/z_Linux_asm.S
         sed -i'' -e 's/__kmp_unnamed_critical_addr/___kmp_unnamed_critical_addr/g' runtime/src/z_Linux_asm.S
+        wget https://github.com/nihui/llvm-project/commit/5c12711f9a21f41bea70566bf15a4026804d6b20.patch
+        patch -p2 -i 5c12711f9a21f41bea70566bf15a4026804d6b20.patch
     - name: openmp-armv7k
       if: steps.cache-openmp.outputs.cache-hit != 'true'
       run: |

From 5d680b855af55194f83b3e7453a0b6940fa64f37 Mon Sep 17 00:00:00 2001
From: nihui <shuizhuyuanluo@126.com>
Date: Mon, 1 Apr 2024 21:47:16 +0800
Subject: [PATCH 10/19] update ios toolchain, ncnn target ilp32

---
 CMakeLists.txt                 | 12 ++++++--
 cmake/ncnn_add_layer.cmake     |  4 +--
 src/CMakeLists.txt             |  4 +--
 toolchains/ios.toolchain.cmake | 53 ++++++++++++++++++++++++----------
 4 files changed, 51 insertions(+), 22 deletions(-)

diff --git a/CMakeLists.txt b/CMakeLists.txt
index 785e2cd39264..a67ab6afb5e4 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -148,7 +148,11 @@ if((IOS AND CMAKE_OSX_ARCHITECTURES MATCHES "arm")
     OR ((CMAKE_CXX_COMPILER_ID MATCHES "MSVC" OR (CMAKE_CXX_COMPILER_ID MATCHES "Clang" AND CMAKE_CXX_SIMULATE_ID MATCHES "MSVC" AND CMAKE_CXX_COMPILER_FRONTEND_VARIANT MATCHES "MSVC")) AND (${CMAKE_GENERATOR_PLATFORM} MATCHES "^(arm|arm64)")))
     set(NCNN_TARGET_ARCH arm)
 
-    if(CMAKE_SIZEOF_VOID_P EQUAL 4)
+    if(APPLE AND CMAKE_OSX_ARCHITECTURES STREQUAL "arm64_32")
+        set(NCNN_TARGET_ILP32 TRUE)
+    endif()
+
+    if(CMAKE_SIZEOF_VOID_P EQUAL 4 AND NOT NCNN_TARGET_ILP32)
         if(CMAKE_CXX_COMPILER_ID MATCHES "MSVC" OR (CMAKE_CXX_COMPILER_ID MATCHES "Clang" AND CMAKE_CXX_SIMULATE_ID MATCHES "MSVC" AND CMAKE_CXX_COMPILER_FRONTEND_VARIANT MATCHES "MSVC"))
             set(CMAKE_REQUIRED_FLAGS "/arch:VFPv4")
             check_cxx_source_compiles("#include <arm_neon.h>\nint main() { float32x4_t _a; float16x4_t _s = vcvt_f16_f32(_a); return 0; }" NCNN_COMPILER_SUPPORT_ARM_VFPV4)
@@ -173,7 +177,7 @@ if((IOS AND CMAKE_OSX_ARCHITECTURES MATCHES "arm")
         endif()
     endif()
 
-    if(CMAKE_SIZEOF_VOID_P EQUAL 8)
+    if(CMAKE_SIZEOF_VOID_P EQUAL 8 OR NCNN_TARGET_ILP32)
         if(CMAKE_CXX_COMPILER_ID MATCHES "MSVC" OR (CMAKE_CXX_COMPILER_ID MATCHES "Clang" AND CMAKE_CXX_SIMULATE_ID MATCHES "MSVC" AND CMAKE_CXX_COMPILER_FRONTEND_VARIANT MATCHES "MSVC"))
             set(CMAKE_REQUIRED_FLAGS "/arch:armv8.0")
             check_cxx_source_compiles("#include <arm_neon.h>\nint main() { float32x4_t _a; float16x4_t _s = vcvt_f16_f32(_a); return 0; }" NCNN_COMPILER_SUPPORT_ARM_VFPV4)
@@ -560,7 +564,9 @@ else()
     endif()
 endif()
 
-if(CMAKE_SIZEOF_VOID_P EQUAL 8)
+if(NCNN_TARGET_ILP32)
+    message(STATUS "Target arch: ${NCNN_TARGET_ARCH} 64bit ilp32")
+elseif(CMAKE_SIZEOF_VOID_P EQUAL 8)
     message(STATUS "Target arch: ${NCNN_TARGET_ARCH} 64bit")
 else()
     message(STATUS "Target arch: ${NCNN_TARGET_ARCH} 32bit")
diff --git a/cmake/ncnn_add_layer.cmake b/cmake/ncnn_add_layer.cmake
index 4eeedb010c7d..a41c52e4ed22 100644
--- a/cmake/ncnn_add_layer.cmake
+++ b/cmake/ncnn_add_layer.cmake
@@ -199,7 +199,7 @@ macro(ncnn_add_layer class)
         endif()
     endif()
 
-    if(NCNN_TARGET_ARCH STREQUAL "arm" AND CMAKE_SIZEOF_VOID_P EQUAL 4)
+    if(NCNN_TARGET_ARCH STREQUAL "arm" AND (CMAKE_SIZEOF_VOID_P EQUAL 4 AND NOT NCNN_TARGET_ILP32))
         if(CMAKE_CXX_COMPILER_ID MATCHES "MSVC" OR (CMAKE_CXX_COMPILER_ID MATCHES "Clang" AND CMAKE_CXX_SIMULATE_ID MATCHES "MSVC" AND CMAKE_CXX_COMPILER_FRONTEND_VARIANT MATCHES "MSVC"))
             if(NCNN_VFPV4)
                 ncnn_add_arch_opt_source(${class} vfpv4 "/arch:VFPv4 /D__ARM_FP=0x0E")
@@ -215,7 +215,7 @@ macro(ncnn_add_layer class)
         endif()
     endif()
 
-    if(NCNN_TARGET_ARCH STREQUAL "arm" AND CMAKE_SIZEOF_VOID_P EQUAL 8)
+    if(NCNN_TARGET_ARCH STREQUAL "arm" AND (CMAKE_SIZEOF_VOID_P EQUAL 8 OR NCNN_TARGET_ILP32))
         if(CMAKE_CXX_COMPILER_ID MATCHES "MSVC" OR (CMAKE_CXX_COMPILER_ID MATCHES "Clang" AND CMAKE_CXX_SIMULATE_ID MATCHES "MSVC" AND CMAKE_CXX_COMPILER_FRONTEND_VARIANT MATCHES "MSVC"))
             if(NCNN_VFPV4)
                 ncnn_add_arch_opt_source(${class} vfpv4 " ")
diff --git a/src/CMakeLists.txt b/src/CMakeLists.txt
index 9ba035b94226..d7dfe19c26e1 100644
--- a/src/CMakeLists.txt
+++ b/src/CMakeLists.txt
@@ -470,7 +470,7 @@ if(NCNN_TARGET_ARCH STREQUAL "x86")
     endif()
 endif()
 
-if(NCNN_TARGET_ARCH STREQUAL "arm" AND CMAKE_SIZEOF_VOID_P EQUAL 4)
+if(NCNN_TARGET_ARCH STREQUAL "arm" AND (CMAKE_SIZEOF_VOID_P EQUAL 4 AND NOT NCNN_TARGET_ILP32))
     if(CMAKE_CXX_COMPILER_ID MATCHES "MSVC" OR (CMAKE_CXX_COMPILER_ID MATCHES "Clang" AND CMAKE_CXX_SIMULATE_ID MATCHES "MSVC" AND CMAKE_CXX_COMPILER_FRONTEND_VARIANT MATCHES "MSVC"))
         # always enable neon for msvc arm
         target_compile_options(ncnn PRIVATE /D__arm__ /D__ARM_NEON)
@@ -493,7 +493,7 @@ if(NCNN_TARGET_ARCH STREQUAL "arm" AND CMAKE_SIZEOF_VOID_P EQUAL 4)
     endif()
 endif()
 
-if(NCNN_TARGET_ARCH STREQUAL "arm" AND CMAKE_SIZEOF_VOID_P EQUAL 8)
+if(NCNN_TARGET_ARCH STREQUAL "arm" AND (CMAKE_SIZEOF_VOID_P EQUAL 8 OR NCNN_TARGET_ILP32))
     if(CMAKE_CXX_COMPILER_ID MATCHES "MSVC" OR (CMAKE_CXX_COMPILER_ID MATCHES "Clang" AND CMAKE_CXX_SIMULATE_ID MATCHES "MSVC" AND CMAKE_CXX_COMPILER_FRONTEND_VARIANT MATCHES "MSVC"))
         # always enable neon and vfpv4 for msvc arm64
         target_compile_options(ncnn PRIVATE /D__arm__ /D__aarch64__ /D__ARM_NEON /D__ARM_FP=0x0E)
diff --git a/toolchains/ios.toolchain.cmake b/toolchains/ios.toolchain.cmake
index 44a6fa5abee6..379014166243 100644
--- a/toolchains/ios.toolchain.cmake
+++ b/toolchains/ios.toolchain.cmake
@@ -70,6 +70,7 @@
 #    WATCHOS = Build for armv7k arm64_32 for watchOS.
 #    WATCHOSCOMBINED = Build for armv7k arm64_32 x86_64 watchOS + watchOS Simulator. Combined into FAT STATIC lib (only supported on 3.14+ of CMake with "-G Xcode" argument in combination with the "cmake --install" CMake build step)
 #    SIMULATOR_WATCHOS = Build for x86_64 for watchOS Simulator.
+#    SIMULATORARM64_WATCHOS = Build for arm64 for watchOS Simulator.
 #    MAC = Build for x86_64 macOS.
 #    MAC_ARM64 = Build for Apple Silicon macOS.
 #    MAC_UNIVERSAL = Combined build for x86_64 and Apple Silicon on macOS.
@@ -87,7 +88,7 @@
 #    CMAKE_OSX_SYSROOT, but can also be manually specified (although this should
 #    not be required).
 #
-# DEPLOYMENT_TARGET: Minimum SDK version to target. Default 2.0 on watchOS and 9.0 on tvOS+iOS
+# DEPLOYMENT_TARGET: Minimum SDK version to target. Default 6.0 on watchOS, 13.0 on tvOS+iOS/iPadOS, 11.0 on macOS, 1.0 on visionOS
 #
 # NAMED_LANGUAGE_SUPPORT:
 #    ON (default) = Will require "enable_language(OBJC) and/or enable_language(OBJCXX)" for full OBJC|OBJCXX support
@@ -113,6 +114,7 @@
 #    SIMULATORARM64_TVOS = arm64
 #    WATCHOS = armv7k arm64_32 (if applicable)
 #    SIMULATOR_WATCHOS = x86_64 (i386 has since long been deprecated)
+#    SIMULATORARM64_WATCHOS = arm64
 #    MAC = x86_64
 #    MAC_ARM64 = arm64
 #    MAC_UNIVERSAL = x86_64 arm64
@@ -158,7 +160,7 @@ set(ENV{_IOS_TOOLCHAIN_HAS_RUN} true)
 list(APPEND _supported_platforms
         "OS" "OS64" "OS64COMBINED" "SIMULATOR" "SIMULATOR64" "SIMULATORARM64" "SIMULATOR64COMBINED"
         "TVOS" "TVOSCOMBINED" "SIMULATOR_TVOS" "SIMULATORARM64_TVOS"
-        "WATCHOS" "WATCHOSCOMBINED" "SIMULATOR_WATCHOS"
+        "WATCHOS" "WATCHOSCOMBINED" "SIMULATOR_WATCHOS" "SIMULATORARM64_WATCHOS"
         "MAC" "MAC_ARM64" "MAC_UNIVERSAL"
         "VISIONOS" "SIMULATOR_VISIONOS" "SIMULATOR64_VISIONOS"
         "MAC_CATALYST" "MAC_CATALYST_ARM64")
@@ -257,10 +259,10 @@ set(NAMED_LANGUAGE_SUPPORT_INT ${NAMED_LANGUAGE_SUPPORT} CACHE BOOL
 if(NOT DEFINED DEPLOYMENT_TARGET)
   if (PLATFORM MATCHES "WATCHOS")
     # Unless specified, SDK version 4.0 is used by default as minimum target version (watchOS).
-    set(DEPLOYMENT_TARGET "4.0")
+    set(DEPLOYMENT_TARGET "6.0")
   elseif(PLATFORM STREQUAL "MAC")
     # Unless specified, SDK version 10.13 (High Sierra) is used by default as the minimum target version (macos).
-    set(DEPLOYMENT_TARGET "10.13")
+    set(DEPLOYMENT_TARGET "11.0")
   elseif(PLATFORM STREQUAL "VISIONOS" OR PLATFORM STREQUAL "SIMULATOR_VISIONOS" OR PLATFORM STREQUAL "SIMULATOR64_VISIONOS")
     # Unless specified, SDK version 1.0 is used by default as minimum target version (visionOS).
     set(DEPLOYMENT_TARGET "1.0")
@@ -275,7 +277,7 @@ if(NOT DEFINED DEPLOYMENT_TARGET)
     set(DEPLOYMENT_TARGET "13.1")
   else()
     # Unless specified, SDK version 11.0 is used by default as the minimum target version (iOS, tvOS).
-    set(DEPLOYMENT_TARGET "11.0")
+    set(DEPLOYMENT_TARGET "13.0")
   endif()
   message(STATUS "[DEFAULTS] Using the default min-version since DEPLOYMENT_TARGET not provided!")
 elseif(DEFINED DEPLOYMENT_TARGET AND PLATFORM MATCHES "^MAC_CATALYST" AND ${DEPLOYMENT_TARGET} VERSION_LESS "13.1")
@@ -481,6 +483,14 @@ elseif(PLATFORM_INT STREQUAL "SIMULATOR_WATCHOS")
   else()
     set(APPLE_TARGET_TRIPLE_INT ${ARCHS_SPLIT}-apple-watchos${DEPLOYMENT_TARGET}-simulator)
   endif()
+elseif(PLATFORM_INT STREQUAL "SIMULATORARM64_WATCHOS")
+  set(SDK_NAME watchsimulator)
+  if(NOT ARCHS)
+    set(ARCHS arm64)
+    set(APPLE_TARGET_TRIPLE_INT arm64-apple-watchos${DEPLOYMENT_TARGET}-simulator)
+  else()
+    set(APPLE_TARGET_TRIPLE_INT ${ARCHS_SPLIT}-apple-watchos${DEPLOYMENT_TARGET}-simulator)
+  endif()
 elseif(PLATFORM_INT STREQUAL "SIMULATOR64_VISIONOS")
   set(SDK_NAME xrsimulator)
   if(NOT ARCHS)
@@ -761,6 +771,7 @@ endif()
 set(CMAKE_XCODE_ATTRIBUTE_CODE_SIGNING_REQUIRED "NO")
 set(CMAKE_SHARED_LIBRARY_PREFIX "lib")
 set(CMAKE_SHARED_LIBRARY_SUFFIX ".dylib")
+set(CMAKE_EXTRA_SHARED_LIBRARY_SUFFIXES ".tbd" ".so")
 set(CMAKE_SHARED_MODULE_PREFIX "lib")
 set(CMAKE_SHARED_MODULE_SUFFIX ".so")
 set(CMAKE_C_COMPILER_ABI ELF)
@@ -816,6 +827,9 @@ elseif(PLATFORM_INT STREQUAL "SIMULATORARM64_TVOS")
   elseif(PLATFORM_INT STREQUAL "SIMULATOR_WATCHOS")
     set(SDK_NAME_VERSION_FLAGS
             "-mwatchos-simulator-version-min=${DEPLOYMENT_TARGET}")
+  elseif(PLATFORM_INT STREQUAL "SIMULATORARM64_WATCHOS")
+    set(SDK_NAME_VERSION_FLAGS
+            "-mwatchos-simulator-version-min=${DEPLOYMENT_TARGET}")
   elseif(PLATFORM_INT STREQUAL "MAC")
     set(SDK_NAME_VERSION_FLAGS
             "-mmacosx-version-min=${DEPLOYMENT_TARGET}")
@@ -826,7 +840,7 @@ elseif(PLATFORM_INT STREQUAL "SIMULATORARM64_TVOS")
   endif()
 elseif(NOT PLATFORM_INT MATCHES "^MAC_CATALYST")
   # Newer versions of CMake sets the version min flags correctly, skip this for Mac Catalyst targets
-  set(CMAKE_OSX_DEPLOYMENT_TARGET ${DEPLOYMENT_TARGET})
+  set(CMAKE_OSX_DEPLOYMENT_TARGET ${DEPLOYMENT_TARGET} CACHE INTERNAL "Minimum OS X deployment version")
 endif()
 
 if(DEFINED APPLE_TARGET_TRIPLE_INT)
@@ -887,35 +901,44 @@ endif()
 if(CMAKE_GENERATOR MATCHES "Xcode")
   message(STATUS "Not setting any manual command-line buildflags, since Xcode is selected as the generator. Modifying the Xcode build-settings directly instead.")
 else()
-  set(CMAKE_C_FLAGS "${C_TARGET_FLAGS} ${APPLE_TARGET_TRIPLE_FLAG} ${SDK_NAME_VERSION_FLAGS} ${OBJC_LEGACY_VARS} ${BITCODE} ${VISIBILITY} ${CMAKE_C_FLAGS}")
+  set(CMAKE_C_FLAGS "${C_TARGET_FLAGS} ${APPLE_TARGET_TRIPLE_FLAG} ${SDK_NAME_VERSION_FLAGS} ${OBJC_LEGACY_VARS} ${BITCODE} ${VISIBILITY} ${CMAKE_C_FLAGS}" CACHE INTERNAL
+     "Flags used by the compiler during all C build types.")
   set(CMAKE_C_FLAGS_DEBUG "-O0 -g ${CMAKE_C_FLAGS_DEBUG}")
   set(CMAKE_C_FLAGS_MINSIZEREL "-DNDEBUG -Os ${CMAKE_C_FLAGS_MINSIZEREL}")
   set(CMAKE_C_FLAGS_RELWITHDEBINFO "-DNDEBUG -O2 -g ${CMAKE_C_FLAGS_RELWITHDEBINFO}")
   set(CMAKE_C_FLAGS_RELEASE "-DNDEBUG -O3 ${CMAKE_C_FLAGS_RELEASE}")
-  set(CMAKE_CXX_FLAGS "${C_TARGET_FLAGS} ${APPLE_TARGET_TRIPLE_FLAG} ${SDK_NAME_VERSION_FLAGS} ${OBJC_LEGACY_VARS} ${BITCODE} ${VISIBILITY} ${CMAKE_CXX_FLAGS}")
+  set(CMAKE_CXX_FLAGS "${C_TARGET_FLAGS} ${APPLE_TARGET_TRIPLE_FLAG} ${SDK_NAME_VERSION_FLAGS} ${OBJC_LEGACY_VARS} ${BITCODE} ${VISIBILITY} ${CMAKE_CXX_FLAGS}" CACHE INTERNAL
+     "Flags used by the compiler during all CXX build types.")
   set(CMAKE_CXX_FLAGS_DEBUG "-O0 -g ${CMAKE_CXX_FLAGS_DEBUG}")
   set(CMAKE_CXX_FLAGS_MINSIZEREL "-DNDEBUG -Os ${CMAKE_CXX_FLAGS_MINSIZEREL}")
   set(CMAKE_CXX_FLAGS_RELWITHDEBINFO "-DNDEBUG -O2 -g ${CMAKE_CXX_FLAGS_RELWITHDEBINFO}")
   set(CMAKE_CXX_FLAGS_RELEASE "-DNDEBUG -O3 ${CMAKE_CXX_FLAGS_RELEASE}")
   if(NAMED_LANGUAGE_SUPPORT_INT)
-    set(CMAKE_OBJC_FLAGS "${C_TARGET_FLAGS} ${APPLE_TARGET_TRIPLE_FLAG} ${SDK_NAME_VERSION_FLAGS} ${BITCODE} ${VISIBILITY} ${FOBJC_ARC} ${OBJC_VARS} ${CMAKE_OBJC_FLAGS}")
+    set(CMAKE_OBJC_FLAGS "${C_TARGET_FLAGS} ${APPLE_TARGET_TRIPLE_FLAG} ${SDK_NAME_VERSION_FLAGS} ${BITCODE} ${VISIBILITY} ${FOBJC_ARC} ${OBJC_VARS} ${CMAKE_OBJC_FLAGS}" CACHE INTERNAL
+     "Flags used by the compiler during all OBJC build types.")
     set(CMAKE_OBJC_FLAGS_DEBUG "-O0 -g ${CMAKE_OBJC_FLAGS_DEBUG}")
     set(CMAKE_OBJC_FLAGS_MINSIZEREL "-DNDEBUG -Os ${CMAKE_OBJC_FLAGS_MINSIZEREL}")
     set(CMAKE_OBJC_FLAGS_RELWITHDEBINFO "-DNDEBUG -O2 -g ${CMAKE_OBJC_FLAGS_RELWITHDEBINFO}")
     set(CMAKE_OBJC_FLAGS_RELEASE "-DNDEBUG -O3 ${CMAKE_OBJC_FLAGS_RELEASE}")
-    set(CMAKE_OBJCXX_FLAGS "${C_TARGET_FLAGS} ${APPLE_TARGET_TRIPLE_FLAG} ${SDK_NAME_VERSION_FLAGS} ${BITCODE} ${VISIBILITY} ${FOBJC_ARC} ${OBJC_VARS} ${CMAKE_OBJCXX_FLAGS}")
+    set(CMAKE_OBJCXX_FLAGS "${C_TARGET_FLAGS} ${APPLE_TARGET_TRIPLE_FLAG} ${SDK_NAME_VERSION_FLAGS} ${BITCODE} ${VISIBILITY} ${FOBJC_ARC} ${OBJC_VARS} ${CMAKE_OBJCXX_FLAGS}" CACHE INTERNAL
+     "Flags used by the compiler during all OBJCXX build types.")
     set(CMAKE_OBJCXX_FLAGS_DEBUG "-O0 -g ${CMAKE_OBJCXX_FLAGS_DEBUG}")
     set(CMAKE_OBJCXX_FLAGS_MINSIZEREL "-DNDEBUG -Os ${CMAKE_OBJCXX_FLAGS_MINSIZEREL}")
     set(CMAKE_OBJCXX_FLAGS_RELWITHDEBINFO "-DNDEBUG -O2 -g ${CMAKE_OBJCXX_FLAGS_RELWITHDEBINFO}")
     set(CMAKE_OBJCXX_FLAGS_RELEASE "-DNDEBUG -O3 ${CMAKE_OBJCXX_FLAGS_RELEASE}")
   endif()
-  set(CMAKE_C_LINK_FLAGS "${C_TARGET_FLAGS} ${SDK_NAME_VERSION_FLAGS} -Wl,-search_paths_first ${CMAKE_C_LINK_FLAGS}")
-  set(CMAKE_CXX_LINK_FLAGS "${C_TARGET_FLAGS} ${SDK_NAME_VERSION_FLAGS}  -Wl,-search_paths_first ${CMAKE_CXX_LINK_FLAGS}")
+  set(CMAKE_C_LINK_FLAGS "${C_TARGET_FLAGS} ${SDK_NAME_VERSION_FLAGS} -Wl,-search_paths_first ${CMAKE_C_LINK_FLAGS}" CACHE INTERNAL
+     "Flags used by the compiler for all C link types.")
+  set(CMAKE_CXX_LINK_FLAGS "${C_TARGET_FLAGS} ${SDK_NAME_VERSION_FLAGS}  -Wl,-search_paths_first ${CMAKE_CXX_LINK_FLAGS}" CACHE INTERNAL
+     "Flags used by the compiler for all CXX link types.")
   if(NAMED_LANGUAGE_SUPPORT_INT)
-    set(CMAKE_OBJC_LINK_FLAGS "${C_TARGET_FLAGS} ${SDK_NAME_VERSION_FLAGS} -Wl,-search_paths_first ${CMAKE_OBJC_LINK_FLAGS}")
-    set(CMAKE_OBJCXX_LINK_FLAGS "${C_TARGET_FLAGS} ${SDK_NAME_VERSION_FLAGS} -Wl,-search_paths_first ${CMAKE_OBJCXX_LINK_FLAGS}")
+    set(CMAKE_OBJC_LINK_FLAGS "${C_TARGET_FLAGS} ${SDK_NAME_VERSION_FLAGS} -Wl,-search_paths_first ${CMAKE_OBJC_LINK_FLAGS}" CACHE INTERNAL
+     "Flags used by the compiler for all OBJC link types.")
+    set(CMAKE_OBJCXX_LINK_FLAGS "${C_TARGET_FLAGS} ${SDK_NAME_VERSION_FLAGS} -Wl,-search_paths_first ${CMAKE_OBJCXX_LINK_FLAGS}" CACHE INTERNAL
+     "Flags used by the compiler for all OBJCXX link types.")
   endif()
-  set(CMAKE_ASM_FLAGS "${CMAKE_C_FLAGS} -x assembler-with-cpp -arch ${CMAKE_OSX_ARCHITECTURES} ${APPLE_TARGET_TRIPLE_FLAG}")
+  set(CMAKE_ASM_FLAGS "${CMAKE_C_FLAGS} -x assembler-with-cpp -arch ${CMAKE_OSX_ARCHITECTURES} ${APPLE_TARGET_TRIPLE_FLAG}" CACHE INTERNAL
+     "Flags used by the compiler for all ASM build types.")
 endif()
 
 ## Print status messages to inform of the current state

From 4ff31e204756b6df75c4fe28aca66d9054b8c7b9 Mon Sep 17 00:00:00 2001
From: nihui <shuizhuyuanluo@126.com>
Date: Mon, 1 Apr 2024 22:09:45 +0800
Subject: [PATCH 11/19] hack

---
 .github/workflows/watchos.yml | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/.github/workflows/watchos.yml b/.github/workflows/watchos.yml
index 57d87f7023aa..6887802a5d9f 100644
--- a/.github/workflows/watchos.yml
+++ b/.github/workflows/watchos.yml
@@ -95,10 +95,14 @@ jobs:
       if: steps.cache-openmp.outputs.cache-hit != 'true'
       run: |
         cd openmp-${{ env.OPENMP_VERSION }}.src
+        wget https://github.com/nihui/llvm-project/commit/82ffbddacd16d4d088e7c0ffccdb58517e20ea4d.patch
+        patch -p2 -i 82ffbddacd16d4d088e7c0ffccdb58517e20ea4d.patch
         mkdir -p build-arm64_32 && cd build-arm64_32
         cmake ${{ env.OPENMP_CMAKE_OPTIONS }} -DPLATFORM=WATCHOS -DARCHS="arm64_32" ..
         cmake --build . -j 4
         cmake --build . --target install
+        cd ..
+        patch -R -p2 -i 82ffbddacd16d4d088e7c0ffccdb58517e20ea4d.patch
     - name: openmp-simulator-x86_64
       if: steps.cache-openmp.outputs.cache-hit != 'true'
       run: |

From 2769637cbed7074c961ecb78712e06226b7d4b55 Mon Sep 17 00:00:00 2001
From: nihui <shuizhuyuanluo@126.com>
Date: Mon, 1 Apr 2024 22:17:26 +0800
Subject: [PATCH 12/19] hack

---
 .github/workflows/watchos.yml | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/.github/workflows/watchos.yml b/.github/workflows/watchos.yml
index 6887802a5d9f..bc255c391c5b 100644
--- a/.github/workflows/watchos.yml
+++ b/.github/workflows/watchos.yml
@@ -95,14 +95,14 @@ jobs:
       if: steps.cache-openmp.outputs.cache-hit != 'true'
       run: |
         cd openmp-${{ env.OPENMP_VERSION }}.src
-        wget https://github.com/nihui/llvm-project/commit/82ffbddacd16d4d088e7c0ffccdb58517e20ea4d.patch
-        patch -p2 -i 82ffbddacd16d4d088e7c0ffccdb58517e20ea4d.patch
+        wget https://github.com/nihui/llvm-project/commit/fd5669f5679d8a44fe59393a0f895d371a76caba.patch
+        patch -p2 -i fd5669f5679d8a44fe59393a0f895d371a76caba.patch
         mkdir -p build-arm64_32 && cd build-arm64_32
         cmake ${{ env.OPENMP_CMAKE_OPTIONS }} -DPLATFORM=WATCHOS -DARCHS="arm64_32" ..
         cmake --build . -j 4
         cmake --build . --target install
         cd ..
-        patch -R -p2 -i 82ffbddacd16d4d088e7c0ffccdb58517e20ea4d.patch
+        patch -R -p2 -i fd5669f5679d8a44fe59393a0f895d371a76caba.patch
     - name: openmp-simulator-x86_64
       if: steps.cache-openmp.outputs.cache-hit != 'true'
       run: |

From ffa014aa80a1494e88108ba3e50e28ded791c185 Mon Sep 17 00:00:00 2001
From: nihui <shuizhuyuanluo@126.com>
Date: Mon, 1 Apr 2024 22:23:26 +0800
Subject: [PATCH 13/19] hack

---
 .github/workflows/watchos.yml | 2 --
 1 file changed, 2 deletions(-)

diff --git a/.github/workflows/watchos.yml b/.github/workflows/watchos.yml
index bc255c391c5b..2564fcdd161c 100644
--- a/.github/workflows/watchos.yml
+++ b/.github/workflows/watchos.yml
@@ -79,8 +79,6 @@ jobs:
         wget https://github.com/llvm/llvm-project/releases/download/llvmorg-${{ env.OPENMP_VERSION }}/openmp-${{ env.OPENMP_VERSION }}.src.tar.xz
         tar -xf openmp-${{ env.OPENMP_VERSION }}.src.tar.xz
         cd openmp-${{ env.OPENMP_VERSION }}.src
-        sed -i'' -e '/.size __kmp_unnamed_critical_addr/d' runtime/src/z_Linux_asm.S
-        sed -i'' -e 's/__kmp_unnamed_critical_addr/___kmp_unnamed_critical_addr/g' runtime/src/z_Linux_asm.S
         wget https://github.com/nihui/llvm-project/commit/5c12711f9a21f41bea70566bf15a4026804d6b20.patch
         patch -p2 -i 5c12711f9a21f41bea70566bf15a4026804d6b20.patch
     - name: openmp-armv7k

From d580a3048ac2caa6ead21a70e3616ffa86ab56b3 Mon Sep 17 00:00:00 2001
From: nihui <shuizhuyuanluo@126.com>
Date: Mon, 1 Apr 2024 22:28:39 +0800
Subject: [PATCH 14/19] hack

---
 .github/workflows/watchos.yml | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/.github/workflows/watchos.yml b/.github/workflows/watchos.yml
index 2564fcdd161c..09d44e80afa0 100644
--- a/.github/workflows/watchos.yml
+++ b/.github/workflows/watchos.yml
@@ -79,6 +79,8 @@ jobs:
         wget https://github.com/llvm/llvm-project/releases/download/llvmorg-${{ env.OPENMP_VERSION }}/openmp-${{ env.OPENMP_VERSION }}.src.tar.xz
         tar -xf openmp-${{ env.OPENMP_VERSION }}.src.tar.xz
         cd openmp-${{ env.OPENMP_VERSION }}.src
+        wget https://github.com/nihui/llvm-project/commit/0f09753bd22af28ad9292aa5c8d2018bc35feb5a.patch
+        patch -p2 -i 0f09753bd22af28ad9292aa5c8d2018bc35feb5a.patch
         wget https://github.com/nihui/llvm-project/commit/5c12711f9a21f41bea70566bf15a4026804d6b20.patch
         patch -p2 -i 5c12711f9a21f41bea70566bf15a4026804d6b20.patch
     - name: openmp-armv7k

From 026f899b682d54541b820a7076c3b66ed96b85aa Mon Sep 17 00:00:00 2001
From: nihuini <nihuini@tencent.com>
Date: Tue, 2 Apr 2024 19:53:46 +0800
Subject: [PATCH 15/19] wip

---
 .github/workflows/visionos.yml | 13 +++++++++----
 .github/workflows/watchos.yml  | 15 +++++++--------
 2 files changed, 16 insertions(+), 12 deletions(-)

diff --git a/.github/workflows/visionos.yml b/.github/workflows/visionos.yml
index 90f4c8092870..1238ab4bca60 100644
--- a/.github/workflows/visionos.yml
+++ b/.github/workflows/visionos.yml
@@ -38,7 +38,7 @@ jobs:
   build:
     runs-on: macos-13
     env:
-      OPENMP_VERSION: '14.0.6'
+      OPENMP_VERSION: '18.1.2'
       OPENMP_CMAKE_OPTIONS: |
         -DCMAKE_TOOLCHAIN_FILE=../../toolchains/ios.toolchain.cmake \
         -DDEPLOYMENT_TARGET=$VISIONOS_DEPLOYMENT_TARGET \
@@ -72,15 +72,20 @@ jobs:
       uses: actions/cache@v4
       with:
         path: openmp-install
-        key: openmp-visionos-install-20240401
+        key: openmp-visionos-install-20240402
     - name: openmp
       if: steps.cache-openmp.outputs.cache-hit != 'true'
       run: |
+        wget https://github.com/llvm/llvm-project/releases/download/llvmorg-${{ env.OPENMP_VERSION }}/cmake-${{ env.OPENMP_VERSION }}.src.tar.xz
+        tar -xf cmake-${{ env.OPENMP_VERSION }}.src.tar.xz
         wget https://github.com/llvm/llvm-project/releases/download/llvmorg-${{ env.OPENMP_VERSION }}/openmp-${{ env.OPENMP_VERSION }}.src.tar.xz
         tar -xf openmp-${{ env.OPENMP_VERSION }}.src.tar.xz
+        mv cmake-${{ env.OPENMP_VERSION }}.src cmake
         cd openmp-${{ env.OPENMP_VERSION }}.src
-        sed -i'' -e '/.size __kmp_unnamed_critical_addr/d' runtime/src/z_Linux_asm.S
-        sed -i'' -e 's/__kmp_unnamed_critical_addr/___kmp_unnamed_critical_addr/g' runtime/src/z_Linux_asm.S
+        wget https://github.com/nihui/llvm-project/commit/ef8c35bcf5d9cfdb0764ffde6a63c04ec715bc37.patch
+        patch -p2 -i ef8c35bcf5d9cfdb0764ffde6a63c04ec715bc37.patch
+        wget https://github.com/nihui/llvm-project/commit/5c12711f9a21f41bea70566bf15a4026804d6b20.patch
+        patch -p2 -i 5c12711f9a21f41bea70566bf15a4026804d6b20.patch
     - name: openmp-arm64
       if: steps.cache-openmp.outputs.cache-hit != 'true'
       run: |
diff --git a/.github/workflows/watchos.yml b/.github/workflows/watchos.yml
index 09d44e80afa0..d67f3c2aa33a 100644
--- a/.github/workflows/watchos.yml
+++ b/.github/workflows/watchos.yml
@@ -38,7 +38,7 @@ jobs:
   build:
     runs-on: macos-13
     env:
-      OPENMP_VERSION: '14.0.6'
+      OPENMP_VERSION: '18.1.2'
       OPENMP_CMAKE_OPTIONS: |
         -DCMAKE_TOOLCHAIN_FILE=../../toolchains/ios.toolchain.cmake \
         -DDEPLOYMENT_TARGET=$WATCHOS_DEPLOYMENT_TARGET \
@@ -72,15 +72,18 @@ jobs:
       uses: actions/cache@v4
       with:
         path: openmp-install
-        key: openmp-watchos-install-20240401
+        key: openmp-watchos-install-20240402
     - name: openmp
       if: steps.cache-openmp.outputs.cache-hit != 'true'
       run: |
+        wget https://github.com/llvm/llvm-project/releases/download/llvmorg-${{ env.OPENMP_VERSION }}/cmake-${{ env.OPENMP_VERSION }}.src.tar.xz
+        tar -xf cmake-${{ env.OPENMP_VERSION }}.src.tar.xz
         wget https://github.com/llvm/llvm-project/releases/download/llvmorg-${{ env.OPENMP_VERSION }}/openmp-${{ env.OPENMP_VERSION }}.src.tar.xz
         tar -xf openmp-${{ env.OPENMP_VERSION }}.src.tar.xz
+        mv cmake-${{ env.OPENMP_VERSION }}.src cmake
         cd openmp-${{ env.OPENMP_VERSION }}.src
-        wget https://github.com/nihui/llvm-project/commit/0f09753bd22af28ad9292aa5c8d2018bc35feb5a.patch
-        patch -p2 -i 0f09753bd22af28ad9292aa5c8d2018bc35feb5a.patch
+        wget https://github.com/nihui/llvm-project/commit/ef8c35bcf5d9cfdb0764ffde6a63c04ec715bc37.patch
+        patch -p2 -i ef8c35bcf5d9cfdb0764ffde6a63c04ec715bc37.patch
         wget https://github.com/nihui/llvm-project/commit/5c12711f9a21f41bea70566bf15a4026804d6b20.patch
         patch -p2 -i 5c12711f9a21f41bea70566bf15a4026804d6b20.patch
     - name: openmp-armv7k
@@ -95,14 +98,10 @@ jobs:
       if: steps.cache-openmp.outputs.cache-hit != 'true'
       run: |
         cd openmp-${{ env.OPENMP_VERSION }}.src
-        wget https://github.com/nihui/llvm-project/commit/fd5669f5679d8a44fe59393a0f895d371a76caba.patch
-        patch -p2 -i fd5669f5679d8a44fe59393a0f895d371a76caba.patch
         mkdir -p build-arm64_32 && cd build-arm64_32
         cmake ${{ env.OPENMP_CMAKE_OPTIONS }} -DPLATFORM=WATCHOS -DARCHS="arm64_32" ..
         cmake --build . -j 4
         cmake --build . --target install
-        cd ..
-        patch -R -p2 -i fd5669f5679d8a44fe59393a0f895d371a76caba.patch
     - name: openmp-simulator-x86_64
       if: steps.cache-openmp.outputs.cache-hit != 'true'
       run: |

From 60f982f2ea09d0fd5a424ca1b44d245a3045cc46 Mon Sep 17 00:00:00 2001
From: nihui <shuizhuyuanluo@126.com>
Date: Tue, 2 Apr 2024 22:48:38 +0800
Subject: [PATCH 16/19] wip

---
 .github/workflows/visionos.yml | 2 +-
 .github/workflows/watchos.yml  | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/.github/workflows/visionos.yml b/.github/workflows/visionos.yml
index 1238ab4bca60..c1da0d9a948a 100644
--- a/.github/workflows/visionos.yml
+++ b/.github/workflows/visionos.yml
@@ -80,7 +80,7 @@ jobs:
         tar -xf cmake-${{ env.OPENMP_VERSION }}.src.tar.xz
         wget https://github.com/llvm/llvm-project/releases/download/llvmorg-${{ env.OPENMP_VERSION }}/openmp-${{ env.OPENMP_VERSION }}.src.tar.xz
         tar -xf openmp-${{ env.OPENMP_VERSION }}.src.tar.xz
-        mv cmake-${{ env.OPENMP_VERSION }}.src cmake
+        mv cmake-${{ env.OPENMP_VERSION }}.src/Modules/* openmp-${{ env.OPENMP_VERSION }}.src/cmake/
         cd openmp-${{ env.OPENMP_VERSION }}.src
         wget https://github.com/nihui/llvm-project/commit/ef8c35bcf5d9cfdb0764ffde6a63c04ec715bc37.patch
         patch -p2 -i ef8c35bcf5d9cfdb0764ffde6a63c04ec715bc37.patch
diff --git a/.github/workflows/watchos.yml b/.github/workflows/watchos.yml
index d67f3c2aa33a..042c6c3ed540 100644
--- a/.github/workflows/watchos.yml
+++ b/.github/workflows/watchos.yml
@@ -80,7 +80,7 @@ jobs:
         tar -xf cmake-${{ env.OPENMP_VERSION }}.src.tar.xz
         wget https://github.com/llvm/llvm-project/releases/download/llvmorg-${{ env.OPENMP_VERSION }}/openmp-${{ env.OPENMP_VERSION }}.src.tar.xz
         tar -xf openmp-${{ env.OPENMP_VERSION }}.src.tar.xz
-        mv cmake-${{ env.OPENMP_VERSION }}.src cmake
+        mv cmake-${{ env.OPENMP_VERSION }}.src/Modules/* openmp-${{ env.OPENMP_VERSION }}.src/cmake/
         cd openmp-${{ env.OPENMP_VERSION }}.src
         wget https://github.com/nihui/llvm-project/commit/ef8c35bcf5d9cfdb0764ffde6a63c04ec715bc37.patch
         patch -p2 -i ef8c35bcf5d9cfdb0764ffde6a63c04ec715bc37.patch

From 59b62c4f3b4f66406f323e74f8d647817e7bc82e Mon Sep 17 00:00:00 2001
From: nihui <shuizhuyuanluo@126.com>
Date: Tue, 2 Apr 2024 22:57:32 +0800
Subject: [PATCH 17/19] wip

---
 .github/workflows/tvos-cpu.yml |  73 ----------------
 .github/workflows/tvos.yml     | 153 +++++++++++++++++++++++++++++++++
 README.md                      |   4 +-
 3 files changed, 155 insertions(+), 75 deletions(-)
 delete mode 100644 .github/workflows/tvos-cpu.yml
 create mode 100644 .github/workflows/tvos.yml

diff --git a/.github/workflows/tvos-cpu.yml b/.github/workflows/tvos-cpu.yml
deleted file mode 100644
index fcfa7097fd6f..000000000000
--- a/.github/workflows/tvos-cpu.yml
+++ /dev/null
@@ -1,73 +0,0 @@
-name: tvos-cpu
-on:
-  push:
-    branches: [master]
-    paths:
-    - '.github/workflows/tvos-cpu.yml'
-    - 'toolchains/ios.toolchain.cmake'
-    - 'CMakeLists.txt'
-    - 'cmake/**'
-    - 'src/*'
-    - 'src/layer/*'
-    - 'src/layer/arm/**'
-  pull_request:
-    branches: [master]
-    paths:
-    - '.github/workflows/tvos-cpu.yml'
-    - 'toolchains/ios.toolchain.cmake'
-    - 'CMakeLists.txt'
-    - 'cmake/**'
-    - 'src/*'
-    - 'src/layer/*'
-    - 'src/layer/arm/**'
-concurrency:
-  group: tvos-cpu-${{ github.ref }}
-  cancel-in-progress: true
-env:
-  DEVELOPER_DIR: /Applications/Xcode_13.4.1.app/Contents/Developer
-  TVOS_DEPLOYMENT_TARGET: '9.0'
-  ENABLE_BITCODE: OFF
-  ENABLE_ARC: OFF
-  ENABLE_VISIBILITY: OFF
-permissions:
-  contents: read
-
-jobs:
-  tvos:
-    runs-on: macos-12
-    steps:
-    - uses: actions/checkout@v4
-
-    - name: build-arm64
-      run: |
-        mkdir build-arm64 && cd build-arm64
-        cmake -DCMAKE_TOOLCHAIN_FILE=../toolchains/ios.toolchain.cmake -DPLATFORM=TVOS -DARCHS="arm64" \
-            -DDEPLOYMENT_TARGET=$TVOS_DEPLOYMENT_TARGET -DENABLE_BITCODE=$ENABLE_BITCODE -DENABLE_ARC=$ENABLE_ARC -DENABLE_VISIBILITY=$ENABLE_VISIBILITY \
-            -DNCNN_OPENMP=OFF -DNCNN_THREADS=OFF \
-            -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=install \
-            ..
-        cmake --build . -j 3
-
-  tvos-simulator:
-    runs-on: macos-12
-    steps:
-    - uses: actions/checkout@v4
-
-    - name: build-x86_64
-      run: |
-        mkdir build-x86_64 && cd build-x86_64
-        cmake -DCMAKE_TOOLCHAIN_FILE=../toolchains/ios.toolchain.cmake -DPLATFORM=SIMULATOR_TVOS -DARCHS="x86_64" \
-            -DDEPLOYMENT_TARGET=$TVOS_DEPLOYMENT_TARGET -DENABLE_BITCODE=$ENABLE_BITCODE -DENABLE_ARC=$ENABLE_ARC -DENABLE_VISIBILITY=$ENABLE_VISIBILITY \
-            -DNCNN_OPENMP=OFF -DNCNN_THREADS=OFF \
-            -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=install \
-            ..
-        cmake --build . -j 3
-    - name: build-arm64
-      run: |
-        mkdir build-arm64 && cd build-arm64
-        cmake -DCMAKE_TOOLCHAIN_FILE=../toolchains/ios.toolchain.cmake -DPLATFORM=SIMULATORARM64_TVOS -DARCHS="arm64" \
-            -DDEPLOYMENT_TARGET=$TVOS_DEPLOYMENT_TARGET -DENABLE_BITCODE=$ENABLE_BITCODE -DENABLE_ARC=$ENABLE_ARC -DENABLE_VISIBILITY=$ENABLE_VISIBILITY \
-            -DNCNN_OPENMP=OFF -DNCNN_THREADS=OFF \
-            -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=install \
-            ..
-        cmake --build . -j 3
diff --git a/.github/workflows/tvos.yml b/.github/workflows/tvos.yml
new file mode 100644
index 000000000000..989ad592fa76
--- /dev/null
+++ b/.github/workflows/tvos.yml
@@ -0,0 +1,153 @@
+name: tvos
+on:
+  push:
+    branches: [master]
+    paths:
+    - '.github/workflows/tvos.yml'
+    - 'toolchains/ios.toolchain.cmake'
+    - 'CMakeLists.txt'
+    - 'cmake/**'
+    - 'src/*'
+    - 'src/layer/*'
+    - 'src/layer/arm/**'
+    - 'src/layer/x86/**'
+  pull_request:
+    branches: [master]
+    paths:
+    - '.github/workflows/tvos.yml'
+    - 'toolchains/ios.toolchain.cmake'
+    - 'CMakeLists.txt'
+    - 'cmake/**'
+    - 'src/*'
+    - 'src/layer/*'
+    - 'src/layer/arm/**'
+    - 'src/layer/x86/**'
+concurrency:
+  group: tvos-${{ github.ref }}
+  cancel-in-progress: true
+env:
+  DEVELOPER_DIR: /Applications/Xcode_15.2.app/Contents/Developer
+  TVOS_DEPLOYMENT_TARGET: '9.0'
+  ENABLE_BITCODE: OFF
+  ENABLE_ARC: OFF
+  ENABLE_VISIBILITY: OFF
+permissions:
+  contents: read
+
+jobs:
+  build:
+    runs-on: macos-13
+    env:
+      OPENMP_VERSION: '18.1.2'
+      OPENMP_CMAKE_OPTIONS: |
+        -DCMAKE_TOOLCHAIN_FILE=../../toolchains/ios.toolchain.cmake \
+        -DDEPLOYMENT_TARGET=$VISIONOS_DEPLOYMENT_TARGET \
+        -DENABLE_BITCODE=$ENABLE_BITCODE \
+        -DENABLE_ARC=$ENABLE_ARC \
+        -DENABLE_VISIBILITY=$ENABLE_VISIBILITY \
+        -DCMAKE_INSTALL_PREFIX=install \
+        -DCMAKE_BUILD_TYPE=Release \
+        -DPERL_EXECUTABLE=/usr/local/bin/perl \
+        -DLIBOMP_ENABLE_SHARED=OFF \
+        -DLIBOMP_OMPT_SUPPORT=OFF \
+        -DLIBOMP_USE_HWLOC=OFF \
+
+      NCNN_CMAKE_OPTIONS: |
+        -DCMAKE_TOOLCHAIN_FILE=../toolchains/ios.toolchain.cmake \
+        -DDEPLOYMENT_TARGET=$VISIONOS_DEPLOYMENT_TARGET \
+        -DENABLE_BITCODE=$ENABLE_BITCODE \
+        -DENABLE_ARC=$ENABLE_ARC \
+        -DENABLE_VISIBILITY=$ENABLE_VISIBILITY \
+        -DCMAKE_INSTALL_PREFIX=install \
+        -DCMAKE_BUILD_TYPE=Release \
+        -DOpenMP_C_FLAGS="-Xclang -fopenmp" -DOpenMP_CXX_FLAGS="-Xclang -fopenmp" \
+        -DOpenMP_C_LIB_NAMES="libomp" -DOpenMP_CXX_LIB_NAMES="libomp" \
+        -DOpenMP_libomp_LIBRARY="libomp.a" \
+
+    steps:
+    - uses: actions/checkout@v4
+
+    - name: cache-openmp
+      id: cache-openmp
+      uses: actions/cache@v4
+      with:
+        path: openmp-install
+        key: openmp-tvos-install-20240402
+    - name: openmp
+      if: steps.cache-openmp.outputs.cache-hit != 'true'
+      run: |
+        wget https://github.com/llvm/llvm-project/releases/download/llvmorg-${{ env.OPENMP_VERSION }}/cmake-${{ env.OPENMP_VERSION }}.src.tar.xz
+        tar -xf cmake-${{ env.OPENMP_VERSION }}.src.tar.xz
+        wget https://github.com/llvm/llvm-project/releases/download/llvmorg-${{ env.OPENMP_VERSION }}/openmp-${{ env.OPENMP_VERSION }}.src.tar.xz
+        tar -xf openmp-${{ env.OPENMP_VERSION }}.src.tar.xz
+        mv cmake-${{ env.OPENMP_VERSION }}.src/Modules/* openmp-${{ env.OPENMP_VERSION }}.src/cmake/
+        cd openmp-${{ env.OPENMP_VERSION }}.src
+        wget https://github.com/nihui/llvm-project/commit/ef8c35bcf5d9cfdb0764ffde6a63c04ec715bc37.patch
+        patch -p2 -i ef8c35bcf5d9cfdb0764ffde6a63c04ec715bc37.patch
+        wget https://github.com/nihui/llvm-project/commit/5c12711f9a21f41bea70566bf15a4026804d6b20.patch
+        patch -p2 -i 5c12711f9a21f41bea70566bf15a4026804d6b20.patch
+    - name: openmp-arm64
+      if: steps.cache-openmp.outputs.cache-hit != 'true'
+      run: |
+        cd openmp-${{ env.OPENMP_VERSION }}.src
+        mkdir -p build-arm64 && cd build-arm64
+        cmake ${{ env.OPENMP_CMAKE_OPTIONS }} -DPLATFORM=TVOS -DARCHS="arm64" ..
+        cmake --build . -j 4
+        cmake --build . --target install
+    - name: openmp-simulator-x86_64
+      if: steps.cache-openmp.outputs.cache-hit != 'true'
+      run: |
+        cd openmp-${{ env.OPENMP_VERSION }}.src
+        mkdir -p build-simulator-x86_64 && cd build-simulator-x86_64
+        cmake ${{ env.OPENMP_CMAKE_OPTIONS }} -DPLATFORM=SIMULATOR_TVOS -DARCHS="x86_64" ..
+        cmake --build . -j 4
+        cmake --build . --target install
+    - name: openmp-simulator-arm64
+      if: steps.cache-openmp.outputs.cache-hit != 'true'
+      run: |
+        cd openmp-${{ env.OPENMP_VERSION }}.src
+        mkdir -p build-simulator-arm64 && cd build-simulator-arm64
+        cmake ${{ env.OPENMP_CMAKE_OPTIONS }} -DPLATFORM=SIMULATOR_TVOS -DARCHS="arm64" ..
+        cmake --build . -j 4
+        cmake --build . --target install
+    - name: openmp-merge-fat-library
+      if: steps.cache-openmp.outputs.cache-hit != 'true'
+      run: |
+        mkdir -p $GITHUB_WORKSPACE/openmp-install
+        mkdir -p $GITHUB_WORKSPACE/openmp-install/tvos
+        mkdir -p $GITHUB_WORKSPACE/openmp-install/tvos-simulator
+
+        cp -a openmp-${{ env.OPENMP_VERSION }}.src/build-arm64/install/include $GITHUB_WORKSPACE/openmp-install/tvos
+        mkdir -p $GITHUB_WORKSPACE/openmp-install/tvos/lib
+        cp openmp-${{ env.OPENMP_VERSION }}.src/build-arm64/install/lib/libomp.a $GITHUB_WORKSPACE/openmp-install/tvos/lib/libomp.a
+
+        cp -a openmp-${{ env.OPENMP_VERSION }}.src/build-simulator-x86_64/install/include $GITHUB_WORKSPACE/openmp-install/tvos-simulator
+        mkdir -p $GITHUB_WORKSPACE/openmp-install/tvos-simulator/lib
+        lipo -create \
+            openmp-${{ env.OPENMP_VERSION }}.src/build-simulator-x86_64/install/lib/libomp.a \
+            openmp-${{ env.OPENMP_VERSION }}.src/build-simulator-arm64/install/lib/libomp.a \
+            -o $GITHUB_WORKSPACE/openmp-install/tvos-simulator/lib/libomp.a
+
+    - name: install-openmp
+      run: |
+        sudo cp $GITHUB_WORKSPACE/openmp-install/tvos/include/* $DEVELOPER_DIR/Platforms/AppleTVOS.platform/Developer/SDKs/AppleTVOS.sdk/usr/include
+        sudo cp $GITHUB_WORKSPACE/openmp-install/tvos/lib/libomp.a $DEVELOPER_DIR/Platforms/AppleTVOS.platform/Developer/SDKs/AppleTVOS.sdk/usr/lib
+
+        sudo cp $GITHUB_WORKSPACE/openmp-install/tvos-simulator/include/* $DEVELOPER_DIR/Platforms/AppleTVSimulator.platform/Developer/SDKs/AppleTVSimulator.sdk/usr/include
+        sudo cp $GITHUB_WORKSPACE/openmp-install/tvos-simulator/lib/libomp.a $DEVELOPER_DIR/Platforms/AppleTVSimulator.platform/Developer/SDKs/AppleTVSimulator.sdk/usr/lib
+
+    - name: arm64
+      run: |
+        mkdir build-arm64 && cd build-arm64
+        cmake ${{ env.NCNN_CMAKE_OPTIONS }} -DPLATFORM=TVOS -DARCHS="arm64" ..
+        cmake --build . -j 4
+    - name: simulator-x86_64
+      run: |
+        mkdir build-x86_64 && cd build-x86_64
+        cmake ${{ env.NCNN_CMAKE_OPTIONS }} -DPLATFORM=SIMULATOR_TVOS -DARCHS="x86_64" ..
+        cmake --build . -j 4
+    - name: simulator-arm64
+      run: |
+        mkdir build-arm64 && cd build-arm64
+        cmake ${{ env.NCNN_CMAKE_OPTIONS }} -DPLATFORM=SIMULATORARM64_TVOS -DARCHS="arm64" ..
+        cmake --build . -j 4
diff --git a/README.md b/README.md
index a4b2876a5e24..be4f7a9b60e1 100644
--- a/README.md
+++ b/README.md
@@ -237,7 +237,7 @@ https://github.com/Tencent/ncnn/releases/latest
 </td>
 <td rowspan=2>
 
-  [<img src="https://img.shields.io/github/actions/workflow/status/Tencent/ncnn/watchos-cpu.yml?branch=master&style=for-the-badge&label=build">](https://github.com/Tencent/ncnn/actions?query=workflow%3Awatchos-cpu)
+  [<img src="https://img.shields.io/github/actions/workflow/status/Tencent/ncnn/watchos.yml?branch=master&style=for-the-badge&label=build">](https://github.com/Tencent/ncnn/actions?query=workflow%3Awatchos)
 
 </td>
 </tr>
@@ -258,7 +258,7 @@ https://github.com/Tencent/ncnn/releases/latest
 </td>
 <td rowspan=2>
 
-  [<img src="https://img.shields.io/github/actions/workflow/status/Tencent/ncnn/tvos-cpu.yml?branch=master&style=for-the-badge&label=build">](https://github.com/Tencent/ncnn/actions?query=workflow%3Atvos-cpu)
+  [<img src="https://img.shields.io/github/actions/workflow/status/Tencent/ncnn/tvos.yml?branch=master&style=for-the-badge&label=build">](https://github.com/Tencent/ncnn/actions?query=workflow%3Atvos)
 
 </td>
 </tr>

From d805697a01ae8c591a01a2241a0644290ee92ec9 Mon Sep 17 00:00:00 2001
From: nihui <shuizhuyuanluo@126.com>
Date: Tue, 2 Apr 2024 23:10:10 +0800
Subject: [PATCH 18/19] wip

---
 .github/workflows/tvos.yml | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/.github/workflows/tvos.yml b/.github/workflows/tvos.yml
index 989ad592fa76..c8ea5111b0dd 100644
--- a/.github/workflows/tvos.yml
+++ b/.github/workflows/tvos.yml
@@ -41,7 +41,7 @@ jobs:
       OPENMP_VERSION: '18.1.2'
       OPENMP_CMAKE_OPTIONS: |
         -DCMAKE_TOOLCHAIN_FILE=../../toolchains/ios.toolchain.cmake \
-        -DDEPLOYMENT_TARGET=$VISIONOS_DEPLOYMENT_TARGET \
+        -DDEPLOYMENT_TARGET=$TVOS_DEPLOYMENT_TARGET \
         -DENABLE_BITCODE=$ENABLE_BITCODE \
         -DENABLE_ARC=$ENABLE_ARC \
         -DENABLE_VISIBILITY=$ENABLE_VISIBILITY \
@@ -54,7 +54,7 @@ jobs:
 
       NCNN_CMAKE_OPTIONS: |
         -DCMAKE_TOOLCHAIN_FILE=../toolchains/ios.toolchain.cmake \
-        -DDEPLOYMENT_TARGET=$VISIONOS_DEPLOYMENT_TARGET \
+        -DDEPLOYMENT_TARGET=$TVOS_DEPLOYMENT_TARGET \
         -DENABLE_BITCODE=$ENABLE_BITCODE \
         -DENABLE_ARC=$ENABLE_ARC \
         -DENABLE_VISIBILITY=$ENABLE_VISIBILITY \

From 0e96bbc60b0eff90cafa79431c6fcc2137d02fda Mon Sep 17 00:00:00 2001
From: nihui <shuizhuyuanluo@126.com>
Date: Tue, 2 Apr 2024 23:22:44 +0800
Subject: [PATCH 19/19] wip

---
 .github/workflows/tvos.yml | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/.github/workflows/tvos.yml b/.github/workflows/tvos.yml
index c8ea5111b0dd..88bc03734d17 100644
--- a/.github/workflows/tvos.yml
+++ b/.github/workflows/tvos.yml
@@ -143,11 +143,11 @@ jobs:
         cmake --build . -j 4
     - name: simulator-x86_64
       run: |
-        mkdir build-x86_64 && cd build-x86_64
+        mkdir build-simulator-x86_64 && cd build-simulator-x86_64
         cmake ${{ env.NCNN_CMAKE_OPTIONS }} -DPLATFORM=SIMULATOR_TVOS -DARCHS="x86_64" ..
         cmake --build . -j 4
     - name: simulator-arm64
       run: |
-        mkdir build-arm64 && cd build-arm64
+        mkdir build-simulator-arm64 && cd build-simulator-arm64
         cmake ${{ env.NCNN_CMAKE_OPTIONS }} -DPLATFORM=SIMULATORARM64_TVOS -DARCHS="arm64" ..
         cmake --build . -j 4