[GitHub] [incubator-tvm] FrozenGene commented on issue #4803: [Frontend] Asymmetric padding of convolution support
FrozenGene commented on issue #4803: [Frontend] Asymmetric padding of convolution support URL: https://github.com/apache/incubator-tvm/pull/4803#issuecomment-581273697 Currently our QNN seems doesn't like passing asymmetric padding directly to it. Change to WIP to indicate I am working on it. I will notice to review when it is complete. Thanks. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-tvm] yidawang commented on a change in pull request #4747: [ThreadPool] Solve ARM BIG.LITTLE heterogeneous multicores
yidawang commented on a change in pull request #4747: [ThreadPool] Solve ARM BIG.LITTLE heterogeneous multicores URL: https://github.com/apache/incubator-tvm/pull/4747#discussion_r373950272 ## File path: src/runtime/threading_backend.cc ## @@ -133,34 +133,42 @@ class ThreadGroup::Impl { #endif } if (exclude_worker0) { // master thread run task -#if defined(__ANDROID__) - SetFullCpuAffinity(); -#else - // if we set TVM_BIND_MASTER_THREAD to be 1, we will bind master thread - // to core 0. - const char* bind_master_thread = getenv("TVM_BIND_MASTER_THREAD"); - if (bind_master_thread && atoi(bind_master_thread) == 1) { -cpu_set_t cpuset; -CPU_ZERO(&cpuset); -if (reverse) { - CPU_SET(sorted_order_[sorted_order_.size() - 1], &cpuset); -} else { - CPU_SET(sorted_order_[0], &cpuset); -} -pthread_setaffinity_np(pthread_self(), sizeof(cpu_set_t), &cpuset); - } - pthread_atfork(nullptr, nullptr, ThreadGroup::Impl::SetFullCpuAffinity); -#endif + // Master thread will have free migration on needed cores. Review comment: I would suggest add one sentence in the comments: Typically, the OS will schedule the master thread to run at core 0, which is idle otherwise, when other works are running. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-tvm] yidawang commented on a change in pull request #4747: [ThreadPool] Solve ARM BIG.LITTLE heterogeneous multicores
yidawang commented on a change in pull request #4747: [ThreadPool] Solve ARM BIG.LITTLE heterogeneous multicores URL: https://github.com/apache/incubator-tvm/pull/4747#discussion_r373950412 ## File path: src/runtime/threading_backend.cc ## @@ -133,34 +133,42 @@ class ThreadGroup::Impl { #endif } if (exclude_worker0) { // master thread run task -#if defined(__ANDROID__) - SetFullCpuAffinity(); -#else - // if we set TVM_BIND_MASTER_THREAD to be 1, we will bind master thread - // to core 0. - const char* bind_master_thread = getenv("TVM_BIND_MASTER_THREAD"); - if (bind_master_thread && atoi(bind_master_thread) == 1) { -cpu_set_t cpuset; -CPU_ZERO(&cpuset); -if (reverse) { - CPU_SET(sorted_order_[sorted_order_.size() - 1], &cpuset); -} else { - CPU_SET(sorted_order_[0], &cpuset); -} -pthread_setaffinity_np(pthread_self(), sizeof(cpu_set_t), &cpuset); - } - pthread_atfork(nullptr, nullptr, ThreadGroup::Impl::SetFullCpuAffinity); -#endif + // Master thread will have free migration on needed cores. + // See the comment inside SetFullCpuAffinity function to get more detail. + SetFullCpuAffinity(reverse); } #endif } - static void SetFullCpuAffinity() { + void SetFullCpuAffinity(bool reverse) { Review comment: This function is to set master thread to full cpu affinity, the function name should indicate that. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-tvm] Coderx7 edited a comment on issue #4802: Building tvm using llvm on windows 10 fails (ninja: build stopped: subcommand failed.)
Coderx7 edited a comment on issue #4802: Building tvm using llvm on windows 10 fails (ninja: build stopped: subcommand failed.) URL: https://github.com/apache/incubator-tvm/issues/4802#issuecomment-581227116 @tqchen Thanks a lot I removed the GCC from PATH and it detected clang/llvm. However it fails again using llvm built from source code as well . > CLANG_~1: error: unsupported option '-fPIC' for target 'x86_64-pc-windows-msvc' Here is the log : ``` PS D:\Codes\tvm_testbed\tvm_llvm\build> cmake .. -G Ninja -- The C compiler identification is Clang 9.0.1 with GNU-like command-line -- The CXX compiler identification is Clang 9.0.1 with GNU-like command-line -- Check for working C compiler: C:/Program Files/LLVM/bin/clang.exe -- Check for working C compiler: C:/Program Files/LLVM/bin/clang.exe -- works -- Detecting C compiler ABI info -- Detecting C compiler ABI info - done -- Detecting C compile features -- Detecting C compile features - done -- Check for working CXX compiler: C:/Program Files/LLVM/bin/clang++.exe -- Check for working CXX compiler: C:/Program Files/LLVM/bin/clang++.exe -- works -- Detecting CXX compiler ABI info -- Detecting CXX compiler ABI info - done -- Detecting CXX compile features -- Detecting CXX compile features - done Build in Debug mode -- Build with RPC support... -- Build with Graph runtime support... -- Build VTA runtime with target: sim -- Found OpenCL: C:/Windows/System32/OpenCL.dll -- Build with OpenCL support -- Use llvm-config=D:/Codes/tvm_testbed/tools/llvm-9.0.1.src.tar/llvm-9.0.1.src/Release/bin/llvm-config.exe -- D:\Codes\tvm_testbed\tools\llvm-9.0.1.src.tar\llvm-9.0.1.src\Release\include -- Found LLVM_INCLUDE_DIRS=D:\Codes\tvm_testbed\tools\llvm-9.0.1.src.tar\llvm-9.0.1.src\Release\include -- Found LLVM_DEFINITIONS= -D_CRT_SECURE_NO_DEPRECATE -D_CRT_SECURE_NO_WARNINGS -D_CRT_NONSTDC_NO_DEPRECATE -D_CRT_NONSTDC_NO_WARNINGS -D_SCL_SECURE_NO_DEPRECATE -D_SCL_SECURE_NO_WARNINGS -DUNICODE -D_UNICODE -D__STDC_CONSTANT_MACROS -D__STDC_FORMAT_MACROS -D__STDC_LIMIT_MACROS -- Found TVM_LLVM_VERSION=90 -- Build with LLVM -- Set TVM_LLVM_VERSION=90 -- Build with contrib.sort -- Build with contrib.hybriddump -- Performing Test SUPPORT_CXX11 -- Performing Test SUPPORT_CXX11 - Failed -- Build with c++11 -- Build with thread support... -- Looking for pthread.h -- Looking for pthread.h - not found -- Found Threads: TRUE -- Configuring done -- Generating done -- Build files have been written to: D:/Codes/tvm_testbed/tvm_llvm/build PS D:\Codes\tvm_testbed\tvm_llvm\build> ninja [1/315] Building CXX object CMakeFiles/tvm.dir/src/relay/op/nn/nn.cc.obj FAILED: CMakeFiles/tvm.dir/src/relay/op/nn/nn.cc.obj C:\PROGRA~1\LLVM\bin\CLANG_~1.EXE -DDMLC_USE_FOPEN64=0 -DNDEBUG -DTVM_LLVM_VERSION=90 -DTVM_THREADPOOL_USE_OPENMP=0 -Dtvm_EXPORTS -I../include -I../3rdparty/dlpack/include -I../3rdparty/dmlc-core/include -I../3rdparty/rang/include -I../3rdparty/compiler-rt -I../3rdparty/picojson -I../vta/include -I"C:/Program Files (x86)/IntelSWTools/system_studio_2020/OpenCL/sdk/include" -ID:/Codes/tvm_testbed/tools/llvm-9.0.1.src.tar/llvm-9.0.1.src/Release/include -I../topi/include -std=c++11 -O0 -g -Wall -fPIC -g -Xclang -gcodeview -O0 -D_DEBUG -D_DLL -D_MT -Xclang --dependent-lib=msvcrtd -D_CRT_SECURE_NO_DEPRECATE -D_CRT_SECURE_NO_WARNINGS -D_CRT_NONSTDC_NO_DEPRECATE -D_CRT_NONSTDC_NO_WARNINGS -D_SCL_SECURE_NO_DEPRECATE -D_SCL_SECURE_NO_WARNINGS -DUNICODE -D_UNICODE -D__STDC_CONSTANT_MACROS -D__STDC_FORMAT_MACROS -D__STDC_LIMIT_MACROS -MD -MT CMakeFiles/tvm.dir/src/relay/op/nn/nn.cc.obj -MF CMakeFiles\tvm.dir\src\relay\op\nn\nn.cc.obj.d -o CMakeFiles/tvm.dir/src/relay/op/nn/nn.cc.obj -c ../src/relay/op/nn/nn.cc CLANG_~1: error: unsupported option '-fPIC' for target 'x86_64-pc-windows-msvc' [2/315] Building CXX object CMakeFiles/tvm.dir/src/relay/op/nn/pooling.cc.obj FAILED: CMakeFiles/tvm.dir/src/relay/op/nn/pooling.cc.obj C:\PROGRA~1\LLVM\bin\CLANG_~1.EXE -DDMLC_USE_FOPEN64=0 -DNDEBUG -DTVM_LLVM_VERSION=90 -DTVM_THREADPOOL_USE_OPENMP=0 -Dtvm_EXPORTS -I../include -I../3rdparty/dlpack/include -I../3rdparty/dmlc-core/include -I../3rdparty/rang/include -I../3rdparty/compiler-rt -I../3rdparty/picojson -I../vta/include -I"C:/Program Files (x86)/IntelSWTools/system_studio_2020/OpenCL/sdk/include" -ID:/Codes/tvm_testbed/tools/llvm-9.0.1.src.tar/llvm-9.0.1.src/Release/include -I../topi/include -std=c++11 -O0 -g -Wall -fPIC -g -Xclang -gcodeview -O0 -D_DEBUG -D_DLL -D_MT -Xclang --dependent-lib=msvcrtd -D_CRT_SECURE_NO_DEPRECATE -D_CRT_SECURE_NO_WARNINGS -D_CRT_NONSTDC_NO_DEPRECATE -D_CRT_NONSTDC_NO_WARNINGS -D_SCL_SECURE_NO_DEPRECATE -D_SCL_SECURE_NO_WARNINGS -DUNICODE -D_UNICODE -D__STDC_CONSTANT_MACROS -D__STDC_FORMAT_MACROS -D__STDC_LIMIT_MACROS -MD -MT CMakeFiles/tvm.
[GitHub] [incubator-tvm] tqchen opened a new pull request #4804: [LINT] Fix -Wextra
tqchen opened a new pull request #4804: [LINT] Fix -Wextra URL: https://github.com/apache/incubator-tvm/pull/4804 This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-tvm] Coderx7 edited a comment on issue #4802: Building tvm using llvm on windows 10 fails (ninja: build stopped: subcommand failed.)
Coderx7 edited a comment on issue #4802: Building tvm using llvm on windows 10 fails (ninja: build stopped: subcommand failed.) URL: https://github.com/apache/incubator-tvm/issues/4802#issuecomment-581227116 @tqchen Thanks a lot I removed the GCC from PATH and it detected clang/llvm. However it fails again using llvm built from source code as well . > CLANG_~1: error: unsupported option '-fPIC' for target 'x86_64-pc-windows-msvc' Here is the log : ``` PS D:\Codes\tvm_testbed\tvm_llvm\build> cmake .. -G Ninja -- The C compiler identification is Clang 9.0.1 with GNU-like command-line -- The CXX compiler identification is Clang 9.0.1 with GNU-like command-line -- Check for working C compiler: C:/Program Files/LLVM/bin/clang.exe -- Check for working C compiler: C:/Program Files/LLVM/bin/clang.exe -- works -- Detecting C compiler ABI info -- Detecting C compiler ABI info - done -- Detecting C compile features -- Detecting C compile features - done -- Check for working CXX compiler: C:/Program Files/LLVM/bin/clang++.exe -- Check for working CXX compiler: C:/Program Files/LLVM/bin/clang++.exe -- works -- Detecting CXX compiler ABI info -- Detecting CXX compiler ABI info - done -- Detecting CXX compile features -- Detecting CXX compile features - done Build in Debug mode -- Build with RPC support... -- Build with Graph runtime support... -- Build VTA runtime with target: sim -- Found OpenCL: C:/Windows/System32/OpenCL.dll -- Build with OpenCL support -- Use llvm-config=D:/Codes/tvm_testbed/tools/llvm-9.0.1.src.tar/llvm-9.0.1.src/Release/bin/llvm-config.exe -- D:\Codes\tvm_testbed\tools\llvm-9.0.1.src.tar\llvm-9.0.1.src\Release\include -- Found LLVM_INCLUDE_DIRS=D:\Codes\tvm_testbed\tools\llvm-9.0.1.src.tar\llvm-9.0.1.src\Release\include -- Found LLVM_DEFINITIONS= -D_CRT_SECURE_NO_DEPRECATE -D_CRT_SECURE_NO_WARNINGS -D_CRT_NONSTDC_NO_DEPRECATE -D_CRT_NONSTDC_NO_WARNINGS -D_SCL_SECURE_NO_DEPRECATE -D_SCL_SECURE_NO_WARNINGS -DUNICODE -D_UNICODE -D__STDC_CONSTANT_MACROS -D__STDC_FORMAT_MACROS -D__STDC_LIMIT_MACROS -- Found TVM_LLVM_VERSION=90 -- Build with LLVM -- Set TVM_LLVM_VERSION=90 -- Build with contrib.sort -- Build with contrib.hybriddump -- Performing Test SUPPORT_CXX11 -- Performing Test SUPPORT_CXX11 - Failed -- Build with c++11 -- Build with thread support... -- Looking for pthread.h -- Looking for pthread.h - not found -- Found Threads: TRUE -- Configuring done -- Generating done -- Build files have been written to: D:/Codes/tvm_testbed/tvm_llvm/build PS D:\Codes\tvm_testbed\tvm_llvm\build> ninja [1/315] Building CXX object CMakeFiles/tvm.dir/src/relay/op/nn/nn.cc.obj FAILED: CMakeFiles/tvm.dir/src/relay/op/nn/nn.cc.obj C:\PROGRA~1\LLVM\bin\CLANG_~1.EXE -DDMLC_USE_FOPEN64=0 -DNDEBUG -DTVM_LLVM_VERSION=90 -DTVM_THREADPOOL_USE_OPENMP=0 -Dtvm_EXPORTS -I../include -I../3rdparty/dlpack/include -I../3rdparty/dmlc-core/include -I../3rdparty/rang/include -I../3rdparty/compiler-rt -I../3rdparty/picojson -I../vta/include -I"C:/Program Files (x86)/IntelSWTools/system_studio_2020/OpenCL/sdk/include" -ID:/Codes/tvm_testbed/tools/llvm-9.0.1.src.tar/llvm-9.0.1.src/Release/include -I../topi/include -std=c++11 -O0 -g -Wall -fPIC -g -Xclang -gcodeview -O0 -D_DEBUG -D_DLL -D_MT -Xclang --dependent-lib=msvcrtd -D_CRT_SECURE_NO_DEPRECATE -D_CRT_SECURE_NO_WARNINGS -D_CRT_NONSTDC_NO_DEPRECATE -D_CRT_NONSTDC_NO_WARNINGS -D_SCL_SECURE_NO_DEPRECATE -D_SCL_SECURE_NO_WARNINGS -DUNICODE -D_UNICODE -D__STDC_CONSTANT_MACROS -D__STDC_FORMAT_MACROS -D__STDC_LIMIT_MACROS -MD -MT CMakeFiles/tvm.dir/src/relay/op/nn/nn.cc.obj -MF CMakeFiles\tvm.dir\src\relay\op\nn\nn.cc.obj.d -o CMakeFiles/tvm.dir/src/relay/op/nn/nn.cc.obj -c ../src/relay/op/nn/nn.cc CLANG_~1: error: unsupported option '-fPIC' for target 'x86_64-pc-windows-msvc' [2/315] Building CXX object CMakeFiles/tvm.dir/src/relay/op/nn/pooling.cc.obj FAILED: CMakeFiles/tvm.dir/src/relay/op/nn/pooling.cc.obj C:\PROGRA~1\LLVM\bin\CLANG_~1.EXE -DDMLC_USE_FOPEN64=0 -DNDEBUG -DTVM_LLVM_VERSION=90 -DTVM_THREADPOOL_USE_OPENMP=0 -Dtvm_EXPORTS -I../include -I../3rdparty/dlpack/include -I../3rdparty/dmlc-core/include -I../3rdparty/rang/include -I../3rdparty/compiler-rt -I../3rdparty/picojson -I../vta/include -I"C:/Program Files (x86)/IntelSWTools/system_studio_2020/OpenCL/sdk/include" -ID:/Codes/tvm_testbed/tools/llvm-9.0.1.src.tar/llvm-9.0.1.src/Release/include -I../topi/include -std=c++11 -O0 -g -Wall -fPIC -g -Xclang -gcodeview -O0 -D_DEBUG -D_DLL -D_MT -Xclang --dependent-lib=msvcrtd -D_CRT_SECURE_NO_DEPRECATE -D_CRT_SECURE_NO_WARNINGS -D_CRT_NONSTDC_NO_DEPRECATE -D_CRT_NONSTDC_NO_WARNINGS -D_SCL_SECURE_NO_DEPRECATE -D_SCL_SECURE_NO_WARNINGS -DUNICODE -D_UNICODE -D__STDC_CONSTANT_MACROS -D__STDC_FORMAT_MACROS -D__STDC_LIMIT_MACROS -MD -MT CMakeFiles/tvm.dir
[GitHub] [incubator-tvm] FrozenGene commented on issue #4794: Change color channel from BGR to RGB for darknet preprocessing
FrozenGene commented on issue #4794: Change color channel from BGR to RGB for darknet preprocessing URL: https://github.com/apache/incubator-tvm/pull/4794#issuecomment-581227501 Thanks @vizero1 @cchung100m This is merged. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[incubator-tvm] branch master updated (00097b1 -> c21d1ee)
This is an automated email from the ASF dual-hosted git repository. zhaowu pushed a change to branch master in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git. from 00097b1 [QNN] Conv2D with dilation support. (#4796) add c21d1ee Change color channel from BGR to RGB for darknet preprocessing (#4794) No new revisions were added by this update. Summary of changes: python/tvm/relay/testing/darknet.py | 1 + 1 file changed, 1 insertion(+)
[GitHub] [incubator-tvm] FrozenGene merged pull request #4794: Change color channel from BGR to RGB for darknet preprocessing
FrozenGene merged pull request #4794: Change color channel from BGR to RGB for darknet preprocessing URL: https://github.com/apache/incubator-tvm/pull/4794 This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-tvm] Coderx7 commented on issue #4802: Building tvm using llvm on windows 10 fails (ninja: build stopped: subcommand failed.)
Coderx7 commented on issue #4802: Building tvm using llvm on windows 10 fails (ninja: build stopped: subcommand failed.) URL: https://github.com/apache/incubator-tvm/issues/4802#issuecomment-581227116 @tqchen Thanks a lot I removed the GCC from PATH and it detected clang/llvm. however it fails again using llvm built from source code as well . here is the log : ``` PS D:\Codes\tvm_testbed\tvm_llvm\build> cmake .. -G Ninja -- The C compiler identification is Clang 9.0.1 with GNU-like command-line -- The CXX compiler identification is Clang 9.0.1 with GNU-like command-line -- Check for working C compiler: C:/Program Files/LLVM/bin/clang.exe -- Check for working C compiler: C:/Program Files/LLVM/bin/clang.exe -- works -- Detecting C compiler ABI info -- Detecting C compiler ABI info - done -- Detecting C compile features -- Detecting C compile features - done -- Check for working CXX compiler: C:/Program Files/LLVM/bin/clang++.exe -- Check for working CXX compiler: C:/Program Files/LLVM/bin/clang++.exe -- works -- Detecting CXX compiler ABI info -- Detecting CXX compiler ABI info - done -- Detecting CXX compile features -- Detecting CXX compile features - done Build in Debug mode -- Build with RPC support... -- Build with Graph runtime support... -- Build VTA runtime with target: sim -- Found OpenCL: C:/Windows/System32/OpenCL.dll -- Build with OpenCL support -- Use llvm-config=D:/Codes/tvm_testbed/tools/llvm-9.0.1.src.tar/llvm-9.0.1.src/Release/bin/llvm-config.exe -- D:\Codes\tvm_testbed\tools\llvm-9.0.1.src.tar\llvm-9.0.1.src\Release\include -- Found LLVM_INCLUDE_DIRS=D:\Codes\tvm_testbed\tools\llvm-9.0.1.src.tar\llvm-9.0.1.src\Release\include -- Found LLVM_DEFINITIONS= -D_CRT_SECURE_NO_DEPRECATE -D_CRT_SECURE_NO_WARNINGS -D_CRT_NONSTDC_NO_DEPRECATE -D_CRT_NONSTDC_NO_WARNINGS -D_SCL_SECURE_NO_DEPRECATE -D_SCL_SECURE_NO_WARNINGS -DUNICODE -D_UNICODE -D__STDC_CONSTANT_MACROS -D__STDC_FORMAT_MACROS -D__STDC_LIMIT_MACROS -- Found TVM_LLVM_VERSION=90 -- Build with LLVM -- Set TVM_LLVM_VERSION=90 -- Build with contrib.sort -- Build with contrib.hybriddump -- Performing Test SUPPORT_CXX11 -- Performing Test SUPPORT_CXX11 - Failed -- Build with c++11 -- Build with thread support... -- Looking for pthread.h -- Looking for pthread.h - not found -- Found Threads: TRUE -- Configuring done -- Generating done -- Build files have been written to: D:/Codes/tvm_testbed/tvm_llvm/build PS D:\Codes\tvm_testbed\tvm_llvm\build> ninja [1/315] Building CXX object CMakeFiles/tvm.dir/src/relay/op/nn/nn.cc.obj FAILED: CMakeFiles/tvm.dir/src/relay/op/nn/nn.cc.obj C:\PROGRA~1\LLVM\bin\CLANG_~1.EXE -DDMLC_USE_FOPEN64=0 -DNDEBUG -DTVM_LLVM_VERSION=90 -DTVM_THREADPOOL_USE_OPENMP=0 -Dtvm_EXPORTS -I../include -I../3rdparty/dlpack/include -I../3rdparty/dmlc-core/include -I../3rdparty/rang/include -I../3rdparty/compiler-rt -I../3rdparty/picojson -I../vta/include -I"C:/Program Files (x86)/IntelSWTools/system_studio_2020/OpenCL/sdk/include" -ID:/Codes/tvm_testbed/tools/llvm-9.0.1.src.tar/llvm-9.0.1.src/Release/include -I../topi/include -std=c++11 -O0 -g -Wall -fPIC -g -Xclang -gcodeview -O0 -D_DEBUG -D_DLL -D_MT -Xclang --dependent-lib=msvcrtd -D_CRT_SECURE_NO_DEPRECATE -D_CRT_SECURE_NO_WARNINGS -D_CRT_NONSTDC_NO_DEPRECATE -D_CRT_NONSTDC_NO_WARNINGS -D_SCL_SECURE_NO_DEPRECATE -D_SCL_SECURE_NO_WARNINGS -DUNICODE -D_UNICODE -D__STDC_CONSTANT_MACROS -D__STDC_FORMAT_MACROS -D__STDC_LIMIT_MACROS -MD -MT CMakeFiles/tvm.dir/src/relay/op/nn/nn.cc.obj -MF CMakeFiles\tvm.dir\src\relay\op\nn\nn.cc.obj.d -o CMakeFiles/tvm.dir/src/relay/op/nn/nn.cc.obj -c ../src/relay/op/nn/nn.cc CLANG_~1: error: unsupported option '-fPIC' for target 'x86_64-pc-windows-msvc' [2/315] Building CXX object CMakeFiles/tvm.dir/src/relay/op/nn/pooling.cc.obj FAILED: CMakeFiles/tvm.dir/src/relay/op/nn/pooling.cc.obj C:\PROGRA~1\LLVM\bin\CLANG_~1.EXE -DDMLC_USE_FOPEN64=0 -DNDEBUG -DTVM_LLVM_VERSION=90 -DTVM_THREADPOOL_USE_OPENMP=0 -Dtvm_EXPORTS -I../include -I../3rdparty/dlpack/include -I../3rdparty/dmlc-core/include -I../3rdparty/rang/include -I../3rdparty/compiler-rt -I../3rdparty/picojson -I../vta/include -I"C:/Program Files (x86)/IntelSWTools/system_studio_2020/OpenCL/sdk/include" -ID:/Codes/tvm_testbed/tools/llvm-9.0.1.src.tar/llvm-9.0.1.src/Release/include -I../topi/include -std=c++11 -O0 -g -Wall -fPIC -g -Xclang -gcodeview -O0 -D_DEBUG -D_DLL -D_MT -Xclang --dependent-lib=msvcrtd -D_CRT_SECURE_NO_DEPRECATE -D_CRT_SECURE_NO_WARNINGS -D_CRT_NONSTDC_NO_DEPRECATE -D_CRT_NONSTDC_NO_WARNINGS -D_SCL_SECURE_NO_DEPRECATE -D_SCL_SECURE_NO_WARNINGS -DUNICODE -D_UNICODE -D__STDC_CONSTANT_MACROS -D__STDC_FORMAT_MACROS -D__STDC_LIMIT_MACROS -MD -MT CMakeFiles/tvm.dir/src/relay/op/nn/pooling.cc.obj -MF CMakeFiles\tvm.dir\src\relay\op\nn\pooling.cc.obj.d -o CMakeFiles/
[GitHub] [incubator-tvm] FrozenGene opened a new pull request #4803: [Frontend] Asymmetric padding of convolution support
FrozenGene opened a new pull request #4803: [Frontend] Asymmetric padding of convolution support URL: https://github.com/apache/incubator-tvm/pull/4803 As we have supported asymmetric padding, we could avoid extra `pad` operator in the frontend now. Our tf frontend has done this. This pr complete this support in the tflite / coreml / keras frontend. Note: This is not conflict with pr https://github.com/apache/incubator-tvm/pull/4787, which could handle it and legalize it in the topi. Like our MXNet frontend could still keep 2D padding if it doesn't have asymmetric padding. However, like tflite / tf / keras / coreml has asymmetric padding, this pr could make us no extra `pad` operator now and could have better performance described in the rfc https://github.com/apache/incubator-tvm/issues/2682 @inadob @optima2005 @wyc-ruiker @anijain2305 @cchung100m Could you help to review this? Thanks. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-tvm] kevinthesun commented on a change in pull request #4797: [AutoTVM] Minor bug fixes in AutoTVM for QNN graphs
kevinthesun commented on a change in pull request #4797: [AutoTVM] Minor bug fixes in AutoTVM for QNN graphs URL: https://github.com/apache/incubator-tvm/pull/4797#discussion_r373913039 ## File path: python/tvm/autotvm/graph_tuner/utils/utils.py ## @@ -73,7 +73,7 @@ def is_boundary_node(node_entry, input_names): # Operators dependent on original layouts. _LAYOUT_FIXED_OP = ["batch_flatten", "transpose", "reshape", "multibox_prior", "multibox_transform_loc", "where", -"non_max_suppression", "strided_slice"] Review comment: Why do we want to change this? This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-tvm] tqchen closed issue #4802: Building tvm using llvm on windows 10 fails (ninja: build stopped: subcommand failed.)
tqchen closed issue #4802: Building tvm using llvm on windows 10 fails (ninja: build stopped: subcommand failed.) URL: https://github.com/apache/incubator-tvm/issues/4802 This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-tvm] tqchen commented on issue #4802: Building tvm using llvm on windows 10 fails (ninja: build stopped: subcommand failed.)
tqchen commented on issue #4802: Building tvm using llvm on windows 10 fails (ninja: build stopped: subcommand failed.) URL: https://github.com/apache/incubator-tvm/issues/4802#issuecomment-581218382 please open a new post in the discuss forum https://discuss.tvm.ai/ We do not support GCC, you will need to build LLVM from source in windows This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-tvm] FrozenGene commented on a change in pull request #4749: [AutoTVM] Support to pass additional compiler options
FrozenGene commented on a change in pull request #4749: [AutoTVM] Support to pass additional compiler options URL: https://github.com/apache/incubator-tvm/pull/4749#discussion_r373905750 ## File path: python/tvm/module.py ## @@ -162,13 +162,20 @@ def export_library(self, f.write(_PackImportsToC(self, is_system_lib)) files.append(path_cc) -if has_c_module: -options = [] -if "options" in kwargs: -opts = kwargs["options"] -options = opts if isinstance(opts, (list, tuple)) else [opts] -opts = options + ["-I" + path for path in find_include_path()] -kwargs.update({'options': opts}) +# Make sure we won't pass {'options': None} or {'options': [None, '-std=c++11', ...]} +# to compiler. We can not prevent users doing the following code: +# f.export_library(path_dso, cc.create_shared, options=None) +# f.export_library(path_dso, cc.create_shared, options=[None, '-std=c++11']) +if "options" in kwargs or has_c_module: +opts = kwargs["options"] if "options" in kwargs else None +options = opts if isinstance(opts, (list, tuple)) else [opts] +options = [opt for opt in options if opt] +if has_c_module: +options = options + ["-I" + path for path in find_include_path()] +if not options: +del kwargs["options"] +else: +kwargs.update({'options': options}) Review comment: Yes. Your way could avoid twice check of `has_c_module`. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-tvm] FrozenGene commented on issue #4796: [QNN] Conv2D with dilation support.
FrozenGene commented on issue #4796: [QNN] Conv2D with dilation support. URL: https://github.com/apache/incubator-tvm/pull/4796#issuecomment-581216587 Thanks @anijain2305 @jackwish This is merged. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-tvm] FrozenGene merged pull request #4796: [QNN] Conv2D with dilation support.
FrozenGene merged pull request #4796: [QNN] Conv2D with dilation support. URL: https://github.com/apache/incubator-tvm/pull/4796 This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[incubator-tvm] branch master updated (9963cf3 -> 00097b1)
This is an automated email from the ASF dual-hosted git repository. zhaowu pushed a change to branch master in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git. from 9963cf3 [QNN] Doc fix on convolution and dequantize (#4799) add 00097b1 [QNN] Conv2D with dilation support. (#4796) No new revisions were added by this update. Summary of changes: src/relay/qnn/op/convolution.cc | 18 +++--- tests/python/relay/test_op_qnn_conv2d.py | 25 - 2 files changed, 35 insertions(+), 8 deletions(-)
[GitHub] [incubator-tvm] FrozenGene commented on issue #4747: [ThreadPool] Solve ARM BIG.LITTLE heterogeneous multicores
FrozenGene commented on issue #4747: [ThreadPool] Solve ARM BIG.LITTLE heterogeneous multicores URL: https://github.com/apache/incubator-tvm/pull/4747#issuecomment-581216205 @yidawang Would you like to review it again? Thanks. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-tvm] vinx13 merged pull request #4799: [QNN] Doc fix on convolution and dequantize
vinx13 merged pull request #4799: [QNN] Doc fix on convolution and dequantize URL: https://github.com/apache/incubator-tvm/pull/4799 This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[incubator-tvm] branch master updated (396095a -> 9963cf3)
This is an automated email from the ASF dual-hosted git repository. wuwei pushed a change to branch master in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git. from 396095a fix #4670: add bias for fc layer (#4801) add 9963cf3 [QNN] Doc fix on convolution and dequantize (#4799) No new revisions were added by this update. Summary of changes: python/tvm/relay/qnn/op/qnn.py | 12 ++-- 1 file changed, 6 insertions(+), 6 deletions(-)
[GitHub] [incubator-tvm] shoubhik commented on issue #4764: [CI] ci-gpu update blockers
shoubhik commented on issue #4764: [CI] ci-gpu update blockers URL: https://github.com/apache/incubator-tvm/issues/4764#issuecomment-581189877 > given that the mkl part poses accuracy problem, i feel it might be a bad idea to rely on it for testing QNN(see also comment about intel dependency). would be great if we can explore generic alternatives that can test QNN. For the parser part, I think we can start by directly checking alpha equivalence of the graph as well as potentially the comparison to a simulated FP32 version. + @anijain2305 @tqchen most of the code for mxnet qnn has been tested for MKLDNN. Some of the formulas used for quantization, dequantization and convolutions used my the reference Mxnet implementations(fake quantization) is not fully and thorougly tested. Also some of the optimizations available in MKLDNN are not implemented in the stock implementaion. So, my suggestion to unblock this PR would be to use the simple feed forward network test case, where we test only the graph generated by QNN. Once we have the fix from MKLDNN as well we can add proper test cases with MKLDNN at that time. @tqchen and @icemelon9 what do you think? This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-tvm] masahi commented on issue #4756: [Docker] Update torch version to 1.4
masahi commented on issue #4756: [Docker] Update torch version to 1.4 URL: https://github.com/apache/incubator-tvm/pull/4756#issuecomment-581182912 @jwfromm have you tried pytorch 1.4 + onnx 1.6? So far I have no problem with this combo. I want to update onnx CI as well if possible. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-tvm] tqchen closed issue #4670: [Relay] Bias missing in final layer of Mobilenet?
tqchen closed issue #4670: [Relay] Bias missing in final layer of Mobilenet? URL: https://github.com/apache/incubator-tvm/issues/4670 This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-tvm] tqchen merged pull request #4801: fix #4670: add bias for fc layer
tqchen merged pull request #4801: fix #4670: add bias for fc layer URL: https://github.com/apache/incubator-tvm/pull/4801 This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[incubator-tvm] branch master updated: fix #4670: add bias for fc layer (#4801)
This is an automated email from the ASF dual-hosted git repository. tqchen pushed a commit to branch master in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git The following commit(s) were added to refs/heads/master by this push: new 396095a fix #4670: add bias for fc layer (#4801) 396095a is described below commit 396095a3e8ad3d15bdc9c52b938d370d4b5ebbf5 Author: kshitij12345 AuthorDate: Mon Feb 3 00:27:12 2020 +0530 fix #4670: add bias for fc layer (#4801) --- python/tvm/relay/testing/mobilenet.py | 2 ++ 1 file changed, 2 insertions(+) diff --git a/python/tvm/relay/testing/mobilenet.py b/python/tvm/relay/testing/mobilenet.py index f76b0c2..1b3ce03 100644 --- a/python/tvm/relay/testing/mobilenet.py +++ b/python/tvm/relay/testing/mobilenet.py @@ -120,7 +120,9 @@ def mobile_net(num_classes=1000, data_shape=(1, 3, 224, 224), pool = relay.nn.global_avg_pool2d(data=body, layout=layout) flatten = relay.nn.batch_flatten(data=pool) weight = relay.var('fc_weight') +bias = relay.var('fc_bias') fc = relay.nn.dense(data=flatten, weight=weight, units=num_classes) +fc = relay.nn.bias_add(fc, bias) softmax = relay.nn.softmax(data=fc) return relay.Function(relay.analysis.free_vars(softmax), softmax)
[GitHub] [incubator-tvm] tqchen commented on issue #4801: fix #4670: add bias for fc layer
tqchen commented on issue #4801: fix #4670: add bias for fc layer URL: https://github.com/apache/incubator-tvm/pull/4801#issuecomment-581165443 Thanks @kshitij12345 This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-tvm] Coderx7 opened a new issue #4802: Building tvm using llvm on windows 10 fails (ninja: build stopped: subcommand failed.)
Coderx7 opened a new issue #4802: Building tvm using llvm on windows 10 fails (ninja: build stopped: subcommand failed.) URL: https://github.com/apache/incubator-tvm/issues/4802 Hi and thank you for this great project. Last night I tried installing `tvm `using `Visual Studio 16 2019` without a hitch and everything went smoothly. but since it seems, tvm needs llvm to work properly ( at least in my case as pointed out [here](https://discuss.tvm.ai/t/tvmerror-check-failed-bf-nullptr-target-llvm-is-not-enabled/5561)), I decided to build `tvm` with `LLVM ON`. I followed the instrcutions at [docs.tvm.ai/install](https://docs.tvm.ai/install/from_source.html) and built llvm from source and then tried building tvm using Ninja. However this is where it fails. upon using the ninja to build it fails with these errors (building with vs2019 also fails as well): ```bash D:\Codes\tvm_testbed\tvm_llvm\build>cmake .. -G Ninja -- The C compiler identification is GNU 5.1.0 -- The CXX compiler identification is GNU 5.1.0 -- Check for working C compiler: C:/TDM-GCC-64/bin/gcc.exe -- Check for working C compiler: C:/TDM-GCC-64/bin/gcc.exe -- works -- Detecting C compiler ABI info -- Detecting C compiler ABI info - done -- Detecting C compile features -- Detecting C compile features - done -- Check for working CXX compiler: C:/TDM-GCC-64/bin/c++.exe -- Check for working CXX compiler: C:/TDM-GCC-64/bin/c++.exe -- works -- Detecting CXX compiler ABI info -- Detecting CXX compiler ABI info - done -- Detecting CXX compile features -- Detecting CXX compile features - done -- Build with RPC support... -- Build with Graph runtime support... -- Build VTA runtime with target: sim -- Found OpenCL: C:/Windows/System32/OpenCL.dll (found version "2.1") -- Build with OpenCL support -- Use llvm-config=D:/Codes/tvm_testbed/tools/llvm-9.0.1.src.tar/llvm-9.0.1.src/Release/bin/llvm-config.exe -- D:\Codes\tvm_testbed\tools\llvm-9.0.1.src.tar\llvm-9.0.1.src\Release\include -- Found LLVM_INCLUDE_DIRS=D:\Codes\tvm_testbed\tools\llvm-9.0.1.src.tar\llvm-9.0.1.src\Release\include -- Found LLVM_DEFINITIONS= -D_CRT_SECURE_NO_DEPRECATE -D_CRT_SECURE_NO_WARNINGS -D_CRT_NONSTDC_NO_DEPRECATE -D_CRT_NONSTDC_NO_WARNINGS -D_SCL_SECURE_NO_DEPRECATE -D_SCL_SECURE_NO_WARNINGS -DUNICODE -D_UNICODE -D__STDC_CONSTANT_MACROS -D__STDC_FORMAT_MACROS -D__STDC_LIMIT_MACROS -- Found TVM_LLVM_VERSION=90 -- Build with LLVM -- Set TVM_LLVM_VERSION=90 -- Build with contrib.sort -- Build with contrib.hybriddump -- Performing Test SUPPORT_CXX11 -- Performing Test SUPPORT_CXX11 - Success -- Build with c++11 -- Build with thread support... -- Looking for pthread.h -- Looking for pthread.h - found -- Performing Test CMAKE_HAVE_LIBC_PTHREAD -- Performing Test CMAKE_HAVE_LIBC_PTHREAD - Success -- Found Threads: TRUE -- Configuring done -- Generating done -- Build files have been written to: D:/Codes/tvm_testbed/tvm_llvm/build D:\Codes\tvm_testbed\tvm_llvm\build>Ninja [1/315] Building CXX object CMakeFiles/vta_fsim.dir/vta/src/device_api.cc.obj ../vta/src/device_api.cc:1:0: warning: -fPIC ignored for target (all code is position independent) /* ^ [2/315] Building CXX object CMakeFiles/vta_fsim.dir/vta/src/runtime.cc.obj ../vta/src/runtime.cc:1:0: warning: -fPIC ignored for target (all code is position independent) /* ^ ../vta/src/runtime.cc: In function 'void* VTABufferAlloc(size_t)': ../vta/src/runtime.cc:1324:7: warning: 'void* VTABufferAlloc(size_t)' redeclared without dllimport attribute: previous dllimport ignored [-Wattributes] void* VTABufferAlloc(size_t size) { ^ ../vta/src/runtime.cc: In function 'void VTABufferFree(void*)': ../vta/src/runtime.cc:1328:6: warning: 'void VTABufferFree(void*)' redeclared without dllimport attribute: previous dllimport ignored [-Wattributes] void VTABufferFree(void* buffer) { ^ ../vta/src/runtime.cc: In function 'void VTABufferCopy(const void*, size_t, void*, size_t, size_t, int)': ../vta/src/runtime.cc:1332:6: warning: 'void VTABufferCopy(const void*, size_t, void*, size_t, size_t, int)' redeclared without dllimport attribute: previous dllimport ignored [-Wattributes] void VTABufferCopy(const void* from, ^ ../vta/src/runtime.cc: In function 'void* VTATLSCommandHandle()': ../vta/src/runtime.cc:1365:18: warning: 'void* VTATLSCommandHandle()' redeclared without dllimport attribute: previous dllimport ignored [-Wattributes] VTACommandHandle VTATLSCommandHandle() { ^ ../vta/src/runtime.cc: In function 'void VTARuntimeShutdown()': ../vta/src/runtime.cc:1369:6: warning: 'void VTARuntimeShutdown()' redeclared without dllimport attribute: previous dllimport ignored [-Wattributes] void VTARuntimeShutdown() { ^ ../vta/src/runtime.cc: In f
[GitHub] [incubator-tvm] kshitij12345 opened a new pull request #4801: fix #4670: add bias for fc layer
kshitij12345 opened a new pull request #4801: fix #4670: add bias for fc layer URL: https://github.com/apache/incubator-tvm/pull/4801 Fixes #4670 @tqchen @zhiics This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services