WASM fixes (Bump emsdk 4.0.23) (#3582)

scottyeager · web-flow · commit a90437cfee0e · 2026-05-09T10:57:31.000+08:00
Now it supports using threads in the WebAssembly.
diff --git a/.github/workflows/npm.yaml b/.github/workflows/npm.yaml
@@ -21,7 +21,7 @@ jobs:
       fail-fast: false
       matrix:
         os: [ubuntu-latest]
-        python-version: ["3.8"]
+        python-version: ["3.10"]
 
     steps:
       - uses: actions/checkout@v4
@@ -37,7 +37,7 @@ jobs:
       - name: Install emsdk
         uses: mymindstorm/setup-emsdk@v14
         with:
-          version: 3.1.51
+          version: 4.0.23
           actions-cache-folder: 'emsdk-cache'
 
       - name: View emsdk version
diff --git a/.github/workflows/test-nodejs.yaml b/.github/workflows/test-nodejs.yaml
@@ -31,7 +31,7 @@ jobs:
       fail-fast: false
       matrix:
         os: [ubuntu-latest] #, macos-latest] #, windows-2022]
-        python-version: ["3.8"]
+        python-version: ["3.10"]
 
     steps:
       - uses: actions/checkout@v4
@@ -52,7 +52,7 @@ jobs:
       - name: Install emsdk
         uses: mymindstorm/setup-emsdk@v14
         with:
-          version: 3.1.51
+          version: 4.0.23
           actions-cache-folder: 'emsdk-cache'
 
       - name: View emsdk version
diff --git a/.github/workflows/wasm-simd-hf-space-en-asr-zipformer.yaml b/.github/workflows/wasm-simd-hf-space-en-asr-zipformer.yaml
@@ -35,7 +35,7 @@ jobs:
       - name: Install emsdk
         uses: mymindstorm/setup-emsdk@v14
         with:
-          version: 3.1.53
+          version: 4.0.23
           actions-cache-folder: 'emsdk-cache'
 
       - name: View emsdk version
diff --git a/.github/workflows/wasm-simd-hf-space-silero-vad.yaml b/.github/workflows/wasm-simd-hf-space-silero-vad.yaml
@@ -35,7 +35,7 @@ jobs:
       - name: Install emsdk
         uses: mymindstorm/setup-emsdk@v14
         with:
-          version: 3.1.53
+          version: 4.0.23
           actions-cache-folder: 'emsdk-cache'
 
       - name: View emsdk version
diff --git a/.github/workflows/wasm-simd-hf-space-speaker-diarization.yaml b/.github/workflows/wasm-simd-hf-space-speaker-diarization.yaml
@@ -36,7 +36,7 @@ jobs:
       - name: Install emsdk
         uses: mymindstorm/setup-emsdk@v14
         with:
-          version: 3.1.53
+          version: 4.0.23
           actions-cache-folder: 'emsdk-cache'
 
       - name: View emsdk version
diff --git a/.github/workflows/wasm-simd-hf-space-speech-enhancement-gtcrn.yaml b/.github/workflows/wasm-simd-hf-space-speech-enhancement-gtcrn.yaml
@@ -36,7 +36,7 @@ jobs:
       - name: Install emsdk
         uses: mymindstorm/setup-emsdk@v14
         with:
-          version: 3.1.53
+          version: 4.0.23
           actions-cache-folder: 'emsdk-cache'
 
       - name: View emsdk version
diff --git a/.github/workflows/wasm-simd-hf-space-ten-vad.yaml b/.github/workflows/wasm-simd-hf-space-ten-vad.yaml
@@ -36,7 +36,7 @@ jobs:
       - name: Install emsdk
         uses: mymindstorm/setup-emsdk@v14
         with:
-          version: 3.1.53
+          version: 4.0.23
           actions-cache-folder: 'emsdk-cache'
 
       - name: View emsdk version
diff --git a/.github/workflows/wasm-simd-hf-space-tts.yaml b/.github/workflows/wasm-simd-hf-space-tts.yaml
@@ -43,7 +43,7 @@ jobs:
       - name: Install emsdk
         uses: mymindstorm/setup-emsdk@v14
         with:
-          version: 3.1.53
+          version: 4.0.23
           actions-cache-folder: 'emsdk-cache'
 
       - name: View emsdk version
diff --git a/.github/workflows/wasm-simd-hf-space-vad-asr.yaml b/.github/workflows/wasm-simd-hf-space-vad-asr.yaml
@@ -43,7 +43,7 @@ jobs:
       - name: Install emsdk
         uses: mymindstorm/setup-emsdk@v14
         with:
-          version: 3.1.53
+          version: 4.0.23
           actions-cache-folder: 'emsdk-cache'
 
       - name: View emsdk version
diff --git a/.github/workflows/wasm-simd-hf-space-zh-cantonese-en-asr-paraformer.yaml b/.github/workflows/wasm-simd-hf-space-zh-cantonese-en-asr-paraformer.yaml
@@ -35,7 +35,7 @@ jobs:
       - name: Install emsdk
         uses: mymindstorm/setup-emsdk@v14
         with:
-          version: 3.1.53
+          version: 4.0.23
           actions-cache-folder: 'emsdk-cache'
 
       - name: View emsdk version
diff --git a/.github/workflows/wasm-simd-hf-space-zh-en-asr-paraformer.yaml b/.github/workflows/wasm-simd-hf-space-zh-en-asr-paraformer.yaml
@@ -35,7 +35,7 @@ jobs:
       - name: Install emsdk
         uses: mymindstorm/setup-emsdk@v14
         with:
-          version: 3.1.53
+          version: 4.0.23
           actions-cache-folder: 'emsdk-cache'
 
       - name: View emsdk version
diff --git a/.github/workflows/wasm-simd-hf-space-zh-en-asr-zipformer.yaml b/.github/workflows/wasm-simd-hf-space-zh-en-asr-zipformer.yaml
@@ -35,7 +35,7 @@ jobs:
       - name: Install emsdk
         uses: mymindstorm/setup-emsdk@v14
         with:
-          version: 3.1.53
+          version: 4.0.23
           actions-cache-folder: 'emsdk-cache'
 
       - name: View emsdk version
diff --git a/CMakeLists.txt b/CMakeLists.txt
@@ -332,6 +332,12 @@ endif()
 
 if(SHERPA_ONNX_ENABLE_WASM)
   add_definitions(-DSHERPA_ONNX_ENABLE_WASM=1)
+  # Build every wasm artifact with pthread/atomics support so the .a files
+  # can be linked with the threaded ONNX Runtime prebuild and a wasm module
+  # that pre-spawns a worker pool. Browsers require COOP/COEP headers to
+  # expose SharedArrayBuffer; wasm/serve.py is provided for that.
+  string(APPEND CMAKE_C_FLAGS " -pthread")
+  string(APPEND CMAKE_CXX_FLAGS " -pthread")
 endif()
 
 if(SHERPA_ONNX_ENABLE_WASM_KWS)
diff --git a/build-wasm-simd-asr.sh b/build-wasm-simd-asr.sh
@@ -14,8 +14,8 @@ if [ x"$EMSCRIPTEN" == x"" ]; then
     echo "git clone https://github.com/emscripten-core/emsdk.git"
     echo "cd emsdk"
     echo "git pull"
-    echo "./emsdk install 3.1.53"
-    echo "./emsdk activate 3.1.53"
+    echo "./emsdk install 4.0.23"
+    echo "./emsdk activate 4.0.23"
     echo "source ./emsdk_env.sh"
     exit 1
   else
@@ -29,7 +29,7 @@ echo "EMSCRIPTEN: $EMSCRIPTEN"
 if [ ! -f $EMSCRIPTEN/cmake/Modules/Platform/Emscripten.cmake ]; then
   echo "Cannot find $EMSCRIPTEN/cmake/Modules/Platform/Emscripten.cmake"
   echo "Please make sure you have installed emsdk correctly"
-  echo "Hint: emsdk 3.1.53 is known to work. Other versions may not work"
+  echo "Hint: emsdk 4.0.23 is known to work. Other versions may not work"
   exit 1
 fi
 
diff --git a/build-wasm-simd-kws.sh b/build-wasm-simd-kws.sh
@@ -9,8 +9,8 @@ if [ x"$EMSCRIPTEN" == x"" ]; then
     echo "git clone https://github.com/emscripten-core/emsdk.git"
     echo "cd emsdk"
     echo "git pull"
-    echo "./emsdk install 3.1.53"
-    echo "./emsdk activate 3.1.53"
+    echo "./emsdk install 4.0.23"
+    echo "./emsdk activate 4.0.23"
     echo "source ./emsdk_env.sh"
     exit 1
   else
@@ -24,7 +24,7 @@ echo "EMSCRIPTEN: $EMSCRIPTEN"
 if [ ! -f $EMSCRIPTEN/cmake/Modules/Platform/Emscripten.cmake ]; then
   echo "Cannot find $EMSCRIPTEN/cmake/Modules/Platform/Emscripten.cmake"
   echo "Please make sure you have installed emsdk correctly"
-  echo "Hint: emsdk 3.1.53 is known to work. Other versions may not work"
+  echo "Hint: emsdk 4.0.23 is known to work. Other versions may not work"
   exit 1
 fi
 
diff --git a/build-wasm-simd-nodejs.sh b/build-wasm-simd-nodejs.sh
@@ -16,8 +16,8 @@ if [ x"$EMSCRIPTEN" == x"" ]; then
     echo "git clone https://github.com/emscripten-core/emsdk.git"
     echo "cd emsdk"
     echo "git pull"
-    echo "./emsdk install 3.1.53"
-    echo "./emsdk activate 3.1.53"
+    echo "./emsdk install 4.0.23"
+    echo "./emsdk activate 4.0.23"
     echo "source ./emsdk_env.sh"
     exit 1
   else
@@ -31,7 +31,7 @@ echo "EMSCRIPTEN: $EMSCRIPTEN"
 if [ ! -f $EMSCRIPTEN/cmake/Modules/Platform/Emscripten.cmake ]; then
   echo "Cannot find $EMSCRIPTEN/cmake/Modules/Platform/Emscripten.cmake"
   echo "Please make sure you have installed emsdk correctly"
-  echo "Hint: emsdk 3.1.53 is known to work. Other versions may not work"
+  echo "Hint: emsdk 4.0.23 is known to work. Other versions may not work"
   exit 1
 fi
 
diff --git a/build-wasm-simd-speaker-diarization.sh b/build-wasm-simd-speaker-diarization.sh
@@ -14,8 +14,8 @@ if [ x"$EMSCRIPTEN" == x"" ]; then
     echo "git clone https://github.com/emscripten-core/emsdk.git"
     echo "cd emsdk"
     echo "git pull"
-    echo "./emsdk install 3.1.53"
-    echo "./emsdk activate 3.1.53"
+    echo "./emsdk install 4.0.23"
+    echo "./emsdk activate 4.0.23"
     echo "source ./emsdk_env.sh"
     exit 1
   else
@@ -29,7 +29,7 @@ echo "EMSCRIPTEN: $EMSCRIPTEN"
 if [ ! -f $EMSCRIPTEN/cmake/Modules/Platform/Emscripten.cmake ]; then
   echo "Cannot find $EMSCRIPTEN/cmake/Modules/Platform/Emscripten.cmake"
   echo "Please make sure you have installed emsdk correctly"
-  echo "Hint: emsdk 3.1.53 is known to work. Other versions may not work"
+  echo "Hint: emsdk 4.0.23 is known to work. Other versions may not work"
   exit 1
 fi
 
diff --git a/build-wasm-simd-speech-enhancement.sh b/build-wasm-simd-speech-enhancement.sh
@@ -14,8 +14,8 @@ if [ x"$EMSCRIPTEN" == x"" ]; then
     echo "git clone https://github.com/emscripten-core/emsdk.git"
     echo "cd emsdk"
     echo "git pull"
-    echo "./emsdk install 3.1.53"
-    echo "./emsdk activate 3.1.53"
+    echo "./emsdk install 4.0.23"
+    echo "./emsdk activate 4.0.23"
     echo "source ./emsdk_env.sh"
     exit 1
   else
@@ -29,7 +29,7 @@ echo "EMSCRIPTEN: $EMSCRIPTEN"
 if [ ! -f $EMSCRIPTEN/cmake/Modules/Platform/Emscripten.cmake ]; then
   echo "Cannot find $EMSCRIPTEN/cmake/Modules/Platform/Emscripten.cmake"
   echo "Please make sure you have installed emsdk correctly"
-  echo "Hint: emsdk 3.1.53 is known to work. Other versions may not work"
+  echo "Hint: emsdk 4.0.23 is known to work. Other versions may not work"
   exit 1
 fi
 
diff --git a/build-wasm-simd-tts.sh b/build-wasm-simd-tts.sh
@@ -14,8 +14,8 @@ if [ x"$EMSCRIPTEN" == x"" ]; then
     echo "git clone https://github.com/emscripten-core/emsdk.git"
     echo "cd emsdk"
     echo "git pull"
-    echo "./emsdk install 3.1.53"
-    echo "./emsdk activate 3.1.53"
+    echo "./emsdk install 4.0.23"
+    echo "./emsdk activate 4.0.23"
     echo "source ./emsdk_env.sh"
     exit 1
   else
@@ -29,7 +29,7 @@ echo "EMSCRIPTEN: $EMSCRIPTEN"
 if [ ! -f $EMSCRIPTEN/cmake/Modules/Platform/Emscripten.cmake ]; then
   echo "Cannot find $EMSCRIPTEN/cmake/Modules/Platform/Emscripten.cmake"
   echo "Please make sure you have installed emsdk correctly"
-  echo "Hint: emsdk 3.1.53 is known to work. Other versions may not work"
+  echo "Hint: emsdk 4.0.23 is known to work. Other versions may not work"
   exit 1
 fi
 
diff --git a/build-wasm-simd-vad-asr.sh b/build-wasm-simd-vad-asr.sh
@@ -15,8 +15,8 @@ if [ x"$EMSCRIPTEN" == x"" ]; then
     echo "git clone https://github.com/emscripten-core/emsdk.git"
     echo "cd emsdk"
     echo "git pull"
-    echo "./emsdk install 3.1.53"
-    echo "./emsdk activate 3.1.53"
+    echo "./emsdk install 4.0.23"
+    echo "./emsdk activate 4.0.23"
     echo "source ./emsdk_env.sh"
     exit 1
   else
@@ -30,7 +30,7 @@ echo "EMSCRIPTEN: $EMSCRIPTEN"
 if [ ! -f $EMSCRIPTEN/cmake/Modules/Platform/Emscripten.cmake ]; then
   echo "Cannot find $EMSCRIPTEN/cmake/Modules/Platform/Emscripten.cmake"
   echo "Please make sure you have installed emsdk correctly"
-  echo "Hint: emsdk 3.1.53 is known to work. Other versions may not work"
+  echo "Hint: emsdk 4.0.23 is known to work. Other versions may not work"
   exit 1
 fi
 
diff --git a/build-wasm-simd-vad.sh b/build-wasm-simd-vad.sh
@@ -14,8 +14,8 @@ if [ x"$EMSCRIPTEN" == x"" ]; then
     echo "git clone https://github.com/emscripten-core/emsdk.git"
     echo "cd emsdk"
     echo "git pull"
-    echo "./emsdk install 3.1.53"
-    echo "./emsdk activate 3.1.53"
+    echo "./emsdk install 4.0.23"
+    echo "./emsdk activate 4.0.23"
     echo "source ./emsdk_env.sh"
     exit 1
   else
@@ -29,7 +29,7 @@ echo "EMSCRIPTEN: $EMSCRIPTEN"
 if [ ! -f $EMSCRIPTEN/cmake/Modules/Platform/Emscripten.cmake ]; then
   echo "Cannot find $EMSCRIPTEN/cmake/Modules/Platform/Emscripten.cmake"
   echo "Please make sure you have installed emsdk correctly"
-  echo "Hint: emsdk 3.1.53 is known to work. Other versions may not work"
+  echo "Hint: emsdk 4.0.23 is known to work. Other versions may not work"
   exit 1
 fi
 
diff --git a/scripts/nodejs/index.js b/scripts/nodejs/index.js
@@ -1,9 +1,14 @@
 // Copyright (c)  2023-2024  Xiaomi Corporation (authors: Fangjun Kuang)
 'use strict'
 
-const wasmModule = require('./sherpa-onnx-wasm-nodejs.js')();
+// Emscripten >= 3.1.50 made MODULARIZE always return a Promise, even with
+// WASM_ASYNC_COMPILATION=0. The runtime still attaches the exports onto the
+// user-supplied moduleArg synchronously, so we pass our own object and ignore
+// the (already-resolved) Promise — keeping this whole module synchronous.
+const wasmModule = {};
+require('./sherpa-onnx-wasm-nodejs.js')(wasmModule);
 const sherpa_onnx_asr = require('./sherpa-onnx-asr.js');
-const sherpa_onnx_tts = require('./sherpa-onnx-tts.js');
+const sherpa_onnx_tts = require('./sherpa-onnx-tts.node.js');
 const sherpa_onnx_kws = require('./sherpa-onnx-kws.js');
 const sherpa_onnx_wave = require('./sherpa-onnx-wave.js');
 const sherpa_onnx_vad = require('./sherpa-onnx-vad.js');
diff --git a/sherpa-onnx/csrc/session.cc b/sherpa-onnx/csrc/session.cc
@@ -144,7 +144,13 @@ Ort::SessionOptions GetSessionOptionsImpl(
   Ort::SessionOptions sess_opts;
   sess_opts.SetIntraOpNumThreads(num_threads);
 
+#if SHERPA_ONNX_ENABLE_WASM
+  // ORT's wasm-with-pthreads prebuild shares one libc pthread pool between
+  // intra-op and inter-op threads; inter_op > 1 starves the intra-op pool.
+  sess_opts.SetInterOpNumThreads(1);
+#else
   sess_opts.SetInterOpNumThreads(num_threads);
+#endif
 
   std::vector<std::string> available_providers = Ort::GetAvailableProviders();
   std::ostringstream os;
diff --git a/wasm/asr/CMakeLists.txt b/wasm/asr/CMakeLists.txt
@@ -39,6 +39,7 @@ list(JOIN mangled_exported_functions "," all_exported_functions)
 include_directories(${CMAKE_SOURCE_DIR})
 set(MY_FLAGS " -s FORCE_FILESYSTEM=1 -s INITIAL_MEMORY=512MB -s ALLOW_MEMORY_GROWTH=1")
 string(APPEND MY_FLAGS " -sSTACK_SIZE=10485760 ") # 10MB
+string(APPEND MY_FLAGS " -pthread -sPTHREAD_POOL_SIZE=4 ")
 string(APPEND MY_FLAGS " -sEXPORTED_FUNCTIONS=[_CopyHeap,_malloc,_free,${all_exported_functions}] ")
 string(APPEND MY_FLAGS "--preload-file ${CMAKE_CURRENT_SOURCE_DIR}/assets@. ")
 string(APPEND MY_FLAGS " -sEXPORTED_RUNTIME_METHODS=['ccall','stringToUTF8','setValue','getValue','lengthBytesUTF8','UTF8ToString','HEAPU8','HEAP16','HEAP32','HEAPU32','HEAPF32','HEAPF64'] ")
diff --git a/wasm/asr/index.html b/wasm/asr/index.html
@@ -61,7 +61,7 @@ <h3>Description</h3>
     </ul>
     <h3>About This Demo</h3>
     <ul>
-      <li><strong>Private and Secure:</strong> All processing is done locally on your device (CPU) within your browser with a single thread. No server is involved, ensuring privacy and security. You can disconnect from the Internet once this page is loaded.</li>
+      <li><strong>Private and Secure:</strong> All processing is done locally on your device (CPU) within your browser using a multi-threaded ONNX Runtime worker pool. No server is involved, ensuring privacy and security. You can disconnect from the Internet once this page is loaded.</li>
       <li><strong>Efficient Resource Usage:</strong> No GPU is required, leaving system resources available for webLLM analysis.</li>
     </ul>
     <h3>Latest Update</h3>
diff --git a/wasm/kws/CMakeLists.txt b/wasm/kws/CMakeLists.txt
@@ -13,6 +13,7 @@ set(exported_functions
   SherpaOnnxDecodeKeywordStream
   SherpaOnnxDestroyKeywordResult
   SherpaOnnxDestroyKeywordSpotter
+  SherpaOnnxDestroyOnlineStream
   SherpaOnnxGetKeywordResult
   SherpaOnnxIsKeywordStreamReady
   SherpaOnnxOnlineStreamAcceptWaveform
@@ -31,6 +32,7 @@ list(JOIN mangled_exported_functions "," all_exported_functions)
 include_directories(${CMAKE_SOURCE_DIR})
 set(MY_FLAGS "-s FORCE_FILESYSTEM=1 -s INITIAL_MEMORY=512MB -s ALLOW_MEMORY_GROWTH=1")
 string(APPEND MY_FLAGS " -sSTACK_SIZE=10485760 ")
+string(APPEND MY_FLAGS " -pthread -sPTHREAD_POOL_SIZE=4 ")
 string(APPEND MY_FLAGS " -sEXPORTED_FUNCTIONS=[_CopyHeap,_malloc,_free,${all_exported_functions}] ")
 string(APPEND MY_FLAGS "--preload-file ${CMAKE_CURRENT_SOURCE_DIR}/assets@. ")
 string(APPEND MY_FLAGS " -sEXPORTED_RUNTIME_METHODS=['ccall','stringToUTF8','setValue','getValue','lengthBytesUTF8','UTF8ToString','HEAPU8','HEAP16','HEAP32','HEAPU32','HEAPF32','HEAPF64'] ")
diff --git a/wasm/nodejs/CMakeLists.txt b/wasm/nodejs/CMakeLists.txt
@@ -152,7 +152,7 @@ install(TARGETS sherpa-onnx-wasm-nodejs DESTINATION bin/wasm/nodejs)
 install(
   FILES
   ${CMAKE_SOURCE_DIR}/wasm/asr/sherpa-onnx-asr.js
-  ${CMAKE_SOURCE_DIR}/wasm/tts/sherpa-onnx-tts.js
+  ${CMAKE_SOURCE_DIR}/wasm/tts/sherpa-onnx-tts.node.js
   ${CMAKE_SOURCE_DIR}/wasm/kws/sherpa-onnx-kws.js
   ${CMAKE_SOURCE_DIR}/wasm/vad/sherpa-onnx-vad.js
   ${CMAKE_SOURCE_DIR}/wasm/nodejs/sherpa-onnx-punctuation.js
diff --git a/wasm/serve.py b/wasm/serve.py
diff --git a/wasm/speaker-diarization/CMakeLists.txt b/wasm/speaker-diarization/CMakeLists.txt
diff --git a/wasm/speech-enhancement/CMakeLists.txt b/wasm/speech-enhancement/CMakeLists.txt
diff --git a/wasm/tts/CMakeLists.txt b/wasm/tts/CMakeLists.txt
diff --git a/wasm/tts/app-tts.js b/wasm/tts/app-tts.js
diff --git a/wasm/tts/sherpa-onnx-tts.js b/wasm/tts/sherpa-onnx-tts.js
diff --git a/wasm/tts/sherpa-onnx-tts.node.js b/wasm/tts/sherpa-onnx-tts.node.js
diff --git a/wasm/tts/sherpa-onnx-tts.worker.js b/wasm/tts/sherpa-onnx-tts.worker.js
diff --git a/wasm/vad-asr/CMakeLists.txt b/wasm/vad-asr/CMakeLists.txt
diff --git a/wasm/vad-asr/index.html b/wasm/vad-asr/index.html
diff --git a/wasm/vad/CMakeLists.txt b/wasm/vad/CMakeLists.txt
diff --git a/wasm/vad/app-vad.js b/wasm/vad/app-vad.js
diff --git a/wasm/vad/index.html b/wasm/vad/index.html
