Comments (3)
Ubuntu20.04 ndk 18b
git clone xxx.git
mkdir build
cd build
cmake -DCMAKE_TOOLCHAIN_FILE=/home/chendongmin/project/ndk/android-ndk-r18b/build/cmake/android.toolchain.cmake -DANDROID_ABI=arm64-v8a ../
cmake --build . -j
from xnnpack.
You need a newer Android NDK. Clang in NDK r18b doesn't support the intrinsic functions used in some XNNPACK microkernels.
from xnnpack.
When building PyTorch version 1.8.1 with NDK r18, I hit this issue on both the armeabi-v8a and arm64-v8a ABIs. The x86 and x86_64 ABIs were OK.
It looks like NDK r19c supports the needed functions, but I wasn't able to use that to build PyTorch either because of a compiler crash on all ABIs:
1. <eof> parser at end of file
2. Per-file LLVM IR generation
3. aten/src/ATen/native/cpu/UnaryOpsKernel.cpp.DEFAULT.cpp:306:3 <Spelling=aten/src/ATen/native/cpu/UnaryOpsKernel.cpp.DEFAULT.cpp:309:9>: Generating code for declaration 'at::native::(anonymous namespace)::sinc_kernel(at::TensorIterator &)::(anonymous class)::operator()()::(anonymous class)::operator()()::(anonymous class)::operator()'
4. aten/src/ATen/native/cpu/UnaryOpsKernel.cpp.DEFAULT.cpp:306:3 <Spelling=aten/src/ATen/native/cpu/UnaryOpsKernel.cpp.DEFAULT.cpp:312:18>: LLVM IR generation of compound statement ('{}')
#0 0x000000000234c8c6 llvm::sys::PrintStackTrace(llvm::raw_ostream&) (/home/smith/git/chaquo/python/target/toolchains/armeabi-v7a/bin/clang80+++0x234c8c6)
#1 0x000000000234b4d0 llvm::sys::RunSignalHandlers() (/home/smith/git/chaquo/python/target/toolchains/armeabi-v7a/bin/clang80+++0x234b4d0)
#2 0x000000000234cbaa (/home/smith/git/chaquo/python/target/toolchains/armeabi-v7a/bin/clang80+++0x234cbaa)
#3 0x00007f3f2c4ee0c0 __restore_rt (/lib/x86_64-linux-gnu/libpthread.so.0+0x110c0)
#4 0x0000000003638bb1 clang::VarDecl::isThisDeclarationADefinition(clang::ASTContext&) const (/home/smith/git/chaquo/python/target/toolchains/armeabi-v7a/bin/clang80+++0x3638bb1)
#5 0x0000000003638f8b clang::VarDecl::getDefinition(clang::ASTContext&) (/home/smith/git/chaquo/python/target/toolchains/armeabi-v7a/bin/clang80+++0x3638f8b)
#6 0x00000000036396d1 clang::VarDecl::getTemplateInstantiationPattern() const (/home/smith/git/chaquo/python/target/toolchains/armeabi-v7a/bin/clang80+++0x36396d1)
#7 0x00000000025de075 (/home/smith/git/chaquo/python/target/toolchains/armeabi-v7a/bin/clang80+++0x25de075)
#8 0x00000000025de082 (/home/smith/git/chaquo/python/target/toolchains/armeabi-v7a/bin/clang80+++0x25de082)
[repeats]
#255 0x00000000025de082 (/home/smith/git/chaquo/python/target/toolchains/armeabi-v7a/bin/clang80+++0x25de082)
clang80++: error: unable to execute command: Segmentation fault
I then tried NDK r21e, which worked fine.
from xnnpack.
Related Issues (20)
- `xnn_weights_cache_provider` look_up doesn't work? HOT 3
- How can I parallelize the execution of this benchmark? (https://github.com/google/XNNPACK/blob/master/bench/spmm-benchmark.h) HOT 2
- cmake build failure with XNNPACK_BUILD_TESTS=ON and XNNPACK_LIBRARY_TYPE=shared
- test/sigmoid_nc_test fails on Hexagon simulator HOT 1
- Load-from-misaligned-address failures on Hexagon simulator HOT 3
- XNNPACK tests that use mmap() fail on Hexagon devices
- Default condition missing for xnnpack_aggregate_library HOT 1
- unsupported instruction `vpdpbusd' HOT 4
- tests for vbinary f16_vsqrdiffc_test are missing a bazel build target
- Add benchmarks for vbinary microkernels HOT 1
- Possible null pointer dereference in logging
- Dynamic shape support follow-up HOT 1
- Enable WASM build on GitHub Actions HOT 1
- Enable QC8/QS8 GEMM/IGEMM for Wasm relaxed integer dot product instruction on x64
- 4x16s4 fp32-gemm kernel have better performance than default(5x16) kernel for meteor lake HOT 1
- QB4W Development
- Failed to compile XNNPACK on WoA(Windows on ARM) device. HOT 5
- Low XNNPACK speedup for ARM Cortex A73 HOT 2
- `f32-vgelu` for HVX build fails HOT 2
- Int 8 GEMM kernel with Int32 Accumulator
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from xnnpack.