christoph/wamr
1
Fork 0
Code Releases Activity

Implement SIMD latest opcodes and update LLVM to 13.0 (#758)

Browse Source 
Implement the latest SIMD opcodes and update LLVM 13.0,
update the llvm build scripts, update the sample workloads‘ build scripts,
and build customized wasi-sdk to build some workloads.
Also refine the CI rules.

Signed-off-by: Wenyong Huang <wenyong.huang@intel.com>
This commit is contained in:
Wenyong Huang
2021-09-17 19:12:57 +08:00
committed by GitHub
parent 7e60a5db8d
commit 7be0d385a6
82 changed files with 5266 additions and 4698 deletions
Download Patch File Download Diff File Expand all files Collapse all files

11
samples/workload/meshoptimizer/CMakeLists.txt
View File

@ -21,9 +21,10 @@ ExternalProject_Add(codecbench
&& ${CMAKE_COMMAND} -E echo "Applying patch"
&& git apply ${CMAKE_CURRENT_SOURCE_DIR}/codecbench.patch
CONFIGURE_COMMAND ${CMAKE_COMMAND}
-DWASI_SDK_PREFIX=${WASI_SDK_HOME}/wasi-sdk
-DCMAKE_TOOLCHAIN_FILE=${WASI_SDK_HOME}/wasi-sdk/share/cmake/wasi-sdk.cmake
${CMAKE_CURRENT_SOURCE_DIR}/meshoptimizer
BUILD_COMMAND make codecbench.opt
INSTALL_COMMAND ${CMAKE_COMMAND} -E copy ./codecbench.opt.wasm ${CMAKE_CURRENT_SOURCE_DIR}/build/codecbench.wasm
-DWASI_SDK_PREFIX=${WASI_SDK_HOME}
-DCMAKE_TOOLCHAIN_FILE=${WASI_SDK_HOME}/share/cmake/wasi-sdk.cmake
-DCMAKE_SYSROOT=${WASI_SDK_HOME}/share/wasi-sysroot
${CMAKE_CURRENT_SOURCE_DIR}/meshoptimizer
BUILD_COMMAND make codecbench
INSTALL_COMMAND ${CMAKE_COMMAND} -E copy ./codecbench.wasm ${CMAKE_CURRENT_SOURCE_DIR}/build/codecbench.wasm
)

1
samples/workload/meshoptimizer/build_workload.sh Symbolic link
View File

@ -0,0 +1 @@
../docker/build_workload.sh

75
samples/workload/meshoptimizer/codecbench.patch
View File

@ -1,8 +1,8 @@
diff --git a/CMakeLists.txt b/CMakeLists.txt
index ffdb4da..a397427 100644
index f4378ce..9bc104b 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -127,3 +127,43 @@ install(FILES
@@ -129,3 +129,43 @@ install(FILES
${CMAKE_CURRENT_BINARY_DIR}/meshoptimizerConfig.cmake
${CMAKE_CURRENT_BINARY_DIR}/meshoptimizerConfigVersion.cmake
DESTINATION ${CMAKE_INSTALL_LIBDIR}/cmake/meshoptimizer)
@ -46,3 +46,74 @@ index ffdb4da..a397427 100644
+)
+
+add_dependencies(codecbench.opt codecbench)
diff --git a/src/vertexcodec.cpp b/src/vertexcodec.cpp
index 5f3ec20..b79bfad 100644
--- a/src/vertexcodec.cpp
+++ b/src/vertexcodec.cpp
@@ -81,13 +81,13 @@
#endif
#ifdef SIMD_WASM
-#define wasmx_splat_v32x4(v, i) wasm_v32x4_shuffle(v, v, i, i, i, i)
-#define wasmx_unpacklo_v8x16(a, b) wasm_v8x16_shuffle(a, b, 0, 16, 1, 17, 2, 18, 3, 19, 4, 20, 5, 21, 6, 22, 7, 23)
-#define wasmx_unpackhi_v8x16(a, b) wasm_v8x16_shuffle(a, b, 8, 24, 9, 25, 10, 26, 11, 27, 12, 28, 13, 29, 14, 30, 15, 31)
-#define wasmx_unpacklo_v16x8(a, b) wasm_v16x8_shuffle(a, b, 0, 8, 1, 9, 2, 10, 3, 11)
-#define wasmx_unpackhi_v16x8(a, b) wasm_v16x8_shuffle(a, b, 4, 12, 5, 13, 6, 14, 7, 15)
-#define wasmx_unpacklo_v64x2(a, b) wasm_v64x2_shuffle(a, b, 0, 2)
-#define wasmx_unpackhi_v64x2(a, b) wasm_v64x2_shuffle(a, b, 1, 3)
+#define wasmx_splat_v32x4(v, i) wasm_i32x4_shuffle(v, v, i, i, i, i)
+#define wasmx_unpacklo_v8x16(a, b) wasm_i8x16_shuffle(a, b, 0, 16, 1, 17, 2, 18, 3, 19, 4, 20, 5, 21, 6, 22, 7, 23)
+#define wasmx_unpackhi_v8x16(a, b) wasm_i8x16_shuffle(a, b, 8, 24, 9, 25, 10, 26, 11, 27, 12, 28, 13, 29, 14, 30, 15, 31)
+#define wasmx_unpacklo_v16x8(a, b) wasm_i16x8_shuffle(a, b, 0, 8, 1, 9, 2, 10, 3, 11)
+#define wasmx_unpackhi_v16x8(a, b) wasm_i16x8_shuffle(a, b, 4, 12, 5, 13, 6, 14, 7, 15)
+#define wasmx_unpacklo_v64x2(a, b) wasm_i64x2_shuffle(a, b, 0, 2)
+#define wasmx_unpackhi_v64x2(a, b) wasm_i64x2_shuffle(a, b, 1, 3)
#endif
namespace meshopt
@@ -700,7 +700,7 @@ static v128_t decodeShuffleMask(unsigned char mask0, unsigned char mask1)
v128_t sm1 = wasm_v128_load(&kDecodeBytesGroupShuffle[mask1]);
v128_t sm1off = wasm_v128_load(&kDecodeBytesGroupCount[mask0]);
- sm1off = wasm_v8x16_shuffle(sm1off, sm1off, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0);
+ sm1off = wasm_i8x16_shuffle(sm1off, sm1off, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0);
v128_t sm1r = wasm_i8x16_add(sm1, sm1off);
@@ -751,7 +751,7 @@ static const unsigned char* decodeBytesGroupSimd(const unsigned char* data, unsi
v128_t shuf = decodeShuffleMask(mask0, mask1);
- v128_t result = wasm_v128_bitselect(wasm_v8x16_swizzle(rest, shuf), sel, mask);
+ v128_t result = wasm_v128_bitselect(wasm_i8x16_swizzle(rest, shuf), sel, mask);
wasm_v128_store(buffer, result);
@@ -773,7 +773,7 @@ static const unsigned char* decodeBytesGroupSimd(const unsigned char* data, unsi
v128_t shuf = decodeShuffleMask(mask0, mask1);
- v128_t result = wasm_v128_bitselect(wasm_v8x16_swizzle(rest, shuf), sel, mask);
+ v128_t result = wasm_v128_bitselect(wasm_i8x16_swizzle(rest, shuf), sel, mask);
wasm_v128_store(buffer, result);
diff --git a/src/vertexfilter.cpp b/src/vertexfilter.cpp
index 023452c..2374cf7 100644
--- a/src/vertexfilter.cpp
+++ b/src/vertexfilter.cpp
@@ -56,10 +56,10 @@
#endif
#ifdef SIMD_WASM
-#define wasmx_unpacklo_v16x8(a, b) wasm_v16x8_shuffle(a, b, 0, 8, 1, 9, 2, 10, 3, 11)
-#define wasmx_unpackhi_v16x8(a, b) wasm_v16x8_shuffle(a, b, 4, 12, 5, 13, 6, 14, 7, 15)
-#define wasmx_unziplo_v32x4(a, b) wasm_v32x4_shuffle(a, b, 0, 2, 4, 6)
-#define wasmx_unziphi_v32x4(a, b) wasm_v32x4_shuffle(a, b, 1, 3, 5, 7)
+#define wasmx_unpacklo_v16x8(a, b) wasm_i16x8_shuffle(a, b, 0, 8, 1, 9, 2, 10, 3, 11)
+#define wasmx_unpackhi_v16x8(a, b) wasm_i16x8_shuffle(a, b, 4, 12, 5, 13, 6, 14, 7, 15)
+#define wasmx_unziplo_v32x4(a, b) wasm_i32x4_shuffle(a, b, 0, 2, 4, 6)
+#define wasmx_unziphi_v32x4(a, b) wasm_i32x4_shuffle(a, b, 1, 3, 5, 7)
#endif
namespace meshopt

1
samples/workload/meshoptimizer/docker_build.sh
View File

@ -1 +0,0 @@
../docker/docker_build.sh