Skip to content
This repository has been archived by the owner on Nov 17, 2023. It is now read-only.

Fixing issue #17840 #18526

Open
wants to merge 27 commits into
base: master
Choose a base branch
from
Open

Fixing issue #17840 #18526

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
27 commits
Select commit Hold shift + click to select a range
6774474
fix the error message of reshape()
May 7, 2020
29ddefe
Fixing issue #16655 reshape() error message
May 7, 2020
7627eaa
test pr
Jun 9, 2020
087a100
fixing #17840
Jun 9, 2020
75f975b
fixing issue #17840
Jun 9, 2020
aafaa99
Merge remote-tracking branch 'upstream/master' into new_branch
Jun 9, 2020
84c389a
fixing issue #17840
Jun 9, 2020
47ba73f
remove blankspace
Jun 9, 2020
060a3b1
fixing bugs
Jun 9, 2020
46256d4
fixing bugs
Jun 10, 2020
bb8727b
fixing cpu errors
Jun 10, 2020
1e406c7
fixing review issues
Jun 11, 2020
c4dbc64
restoring tvmop tests
Jun 12, 2020
a92f500
Apply suggestions from code review
jinboci Jun 12, 2020
beca3b8
Update c_api.cc
jinboci Jun 12, 2020
2bdeb6b
Apply suggestions from code review
Jun 12, 2020
d55a005
Merge branch 'new_branch' of https://github.com/JinboCi/incubator-mxn…
Jun 12, 2020
3183a90
Apply suggestions from code review
Jun 12, 2020
3bb193c
Merge branch 'test_tvmop' into new_branch
Jun 12, 2020
dfd98a4
Update multiarray.py
jinboci Jun 12, 2020
a578b0d
Merge branch 'master' into new_branch
jinboci Jun 15, 2020
ebfc059
fixing bugs
Jun 15, 2020
c75ed75
Update c_api.cc
jinboci Jun 16, 2020
0ee8126
try fixing libcuda.so.1 cannot open shared object file
Jun 16, 2020
4c6298f
reset changes of utls.py
jinboci Jun 18, 2020
da96721
fixing bugs
Jun 18, 2020
165f697
Merge branch 'new_branch' of https://github.com/JinboCi/incubator-mxn…
Jun 18, 2020
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
46 changes: 46 additions & 0 deletions ci/docker/runtime_functions.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -719,6 +719,7 @@ build_ubuntu_gpu_mkldnn() {
CC=gcc-7 CXX=g++-7 cmake \
-DCMAKE_BUILD_TYPE="RelWithDebInfo" \
-DUSE_MKL_IF_AVAILABLE=OFF \
-DUSE_TVM_OP=ON \
-DUSE_CUDA=ON \
-DMXNET_CUDA_ARCH="$CI_CMAKE_CUDA_ARCH" \
-DUSE_CPP_PACKAGE=ON \
Expand All @@ -732,6 +733,7 @@ build_ubuntu_gpu_mkldnn_nocudnn() {
CC=gcc-7 CXX=g++-7 cmake \
-DCMAKE_BUILD_TYPE="RelWithDebInfo" \
-DUSE_MKL_IF_AVAILABLE=OFF \
-DUSE_TVM_OP=ON \
-DUSE_CUDA=ON \
-DMXNET_CUDA_ARCH="$CI_CMAKE_CUDA_ARCH" \
-DUSE_CUDNN=OFF \
Expand All @@ -746,6 +748,7 @@ build_ubuntu_gpu_cuda101_cudnn7() {
CC=gcc-7 CXX=g++-7 cmake \
-DCMAKE_BUILD_TYPE="RelWithDebInfo" \
-DUSE_MKL_IF_AVAILABLE=OFF \
-DUSE_TVM_OP=ON \
-DUSE_CUDA=ON \
-DMXNET_CUDA_ARCH="$CI_CMAKE_CUDA_ARCH" \
-DUSE_CUDNN=ON \
Expand Down Expand Up @@ -785,6 +788,7 @@ build_ubuntu_gpu_cuda101_cudnn7_make() {
USE_CUDA=1 \
USE_CUDA_PATH=/usr/local/cuda \
USE_CUDNN=1 \
USE_TVM_OP=1 \
USE_CPP_PACKAGE=1 \
USE_DIST_KVSTORE=1 \
CUDA_ARCH="$CI_CUDA_COMPUTE_CAPABILITIES" \
Expand All @@ -804,6 +808,7 @@ build_ubuntu_gpu_cuda101_cudnn7_mkldnn_cpp_test() {
USE_CUDA=1 \
USE_CUDA_PATH=/usr/local/cuda \
USE_CUDNN=1 \
USE_TVM_OP=0 \
USE_CPP_PACKAGE=1 \
USE_DIST_KVSTORE=1 \
CUDA_ARCH="$CI_CUDA_COMPUTE_CAPABILITIES" \
Expand All @@ -813,6 +818,23 @@ build_ubuntu_gpu_cuda101_cudnn7_mkldnn_cpp_test() {
make cython PYTHON=python3
}

build_ubuntu_gpu_cuda101_cudnn7_no_tvm_op() {
set -ex
cd /work/build
CC=gcc-7 CXX=g++-7 cmake \
-DCMAKE_BUILD_TYPE="RelWithDebInfo" \
-DUSE_MKL_IF_AVAILABLE=OFF \
-DUSE_TVM_OP=OFF \
-DUSE_CUDA=ON \
-DMXNET_CUDA_ARCH="$CI_CMAKE_CUDA_ARCH" \
-DUSE_CUDNN=ON \
-DUSE_MKLDNN=OFF \
-DBUILD_CYTHON_MODULES=ON \
-DUSE_DIST_KVSTORE=ON \
-G Ninja /work/mxnet
ninja
}

build_ubuntu_amalgamation() {
set -ex
# Amalgamation can not be run with -j nproc
Expand Down Expand Up @@ -843,6 +865,7 @@ build_ubuntu_gpu_cmake() {
-DUSE_SIGNAL_HANDLER=ON \
-DUSE_CUDA=ON \
-DUSE_CUDNN=ON \
-DUSE_TVM_OP=ON \
-DUSE_MKL_IF_AVAILABLE=OFF \
-DUSE_MKLML_MKL=OFF \
-DUSE_MKLDNN=OFF \
Expand All @@ -863,6 +886,7 @@ build_ubuntu_gpu_cmake_no_rtc() {
-DUSE_SIGNAL_HANDLER=ON \
-DUSE_CUDA=ON \
-DUSE_CUDNN=ON \
-DUSE_TVM_OP=ON \
-DUSE_MKL_IF_AVAILABLE=OFF \
-DUSE_MKLML_MKL=OFF \
-DUSE_MKLDNN=ON \
Expand All @@ -877,6 +901,27 @@ build_ubuntu_gpu_cmake_no_rtc() {
ninja
}

build_ubuntu_gpu_cmake_no_tvm_op() {
set -ex
cd /work/build
CC=gcc-7 CXX=g++-7 cmake \
-DUSE_SIGNAL_HANDLER=ON \
-DUSE_CUDA=ON \
-DUSE_CUDNN=ON \
-DUSE_TVM_OP=OFF \
-DUSE_MKL_IF_AVAILABLE=OFF \
-DUSE_MKLML_MKL=OFF \
-DUSE_MKLDNN=OFF \
-DUSE_DIST_KVSTORE=ON \
-DCMAKE_BUILD_TYPE=Release \
-DMXNET_CUDA_ARCH="$CI_CMAKE_CUDA_ARCH" \
-DBUILD_CYTHON_MODULES=1 \
-G Ninja \
/work/mxnet

ninja
}

build_ubuntu_cpu_large_tensor() {
set -ex
cd /work/build
Expand All @@ -899,6 +944,7 @@ build_ubuntu_gpu_large_tensor() {
-DUSE_SIGNAL_HANDLER=ON \
-DUSE_CUDA=ON \
-DUSE_CUDNN=ON \
-DUSE_TVM_OP=ON \
-DUSE_MKL_IF_AVAILABLE=OFF \
-DUSE_MKLML_MKL=OFF \
-DUSE_MKLDNN=OFF \
Expand Down
44 changes: 44 additions & 0 deletions ci/jenkins/Jenkins_steps.groovy
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -310,6 +310,20 @@ def compile_unix_full_gpu_mkldnn_cpp_test(lib_name) {
}]
}

def compile_unix_full_gpu_no_tvm_op(lib_name) {
return ['GPU: CUDA10.1+cuDNN7 TVM_OP OFF': {
node(NODE_LINUX_CPU) {
ws('workspace/build-gpu-no-tvm-op') {
timeout(time: max_time, unit: 'MINUTES') {
utils.init_git()
utils.docker_run('ubuntu_build_cuda', 'build_ubuntu_gpu_cuda101_cudnn7_no_tvm_op', false)
utils.pack_lib(lib_name, mx_lib_cpp_examples_no_tvm_op)
}
}
}
}]
}

def compile_unix_cmake_gpu(lib_name) {
return ['GPU: CMake': {
node(NODE_LINUX_CPU) {
Expand All @@ -324,6 +338,20 @@ def compile_unix_cmake_gpu(lib_name) {
}]
}

def compile_unix_cmake_gpu_no_tvm_op(lib_name) {
return ['GPU: CMake TVM_OP OFF': {
node(NODE_LINUX_CPU) {
ws('workspace/build-cmake-gpu-no-tvm-op') {
timeout(time: max_time, unit: 'MINUTES') {
utils.init_git()
utils.docker_run('ubuntu_gpu_cu101', 'build_ubuntu_gpu_cmake_no_tvm_op', false)
utils.pack_lib(lib_name, mx_cmake_lib_no_tvm_op)
}
}
}
}]
}

def compile_unix_cmake_gpu_no_rtc(lib_name) {
return ['GPU: CMake CUDA RTC OFF': {
node(NODE_LINUX_CPU) {
Expand All @@ -338,6 +366,22 @@ def compile_unix_cmake_gpu_no_rtc(lib_name) {
}]
}

def test_unix_python3_gpu_no_tvm_op() {
return ['Python3: GPU TVM_OP OFF': {
node(NODE_LINUX_GPU) {
ws('workspace/ut-python3-gpu-no-tvm-op') {
try {
utils.unpack_and_init('gpu_no_tvm_op', mx_lib_cpp_examples_no_tvm_op)
python3_gpu_ut_cython('ubuntu_gpu_cu101')
utils.publish_test_coverage()
} finally {
utils.collect_test_results_unix('tests_gpu.xml', 'tests_python3_gpu.xml')
}
}
}
}]
}

def compile_unix_tensorrt_gpu(lib_name) {
return ['TensorRT': {
node(NODE_LINUX_CPU) {
Expand Down
4 changes: 4 additions & 0 deletions ci/jenkins/Jenkinsfile_unix_gpu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -42,6 +42,8 @@ core_logic: {
custom_steps.compile_unix_cmake_gpu('cmake_gpu'),
custom_steps.compile_unix_tensorrt_gpu('tensorrt'),
custom_steps.compile_unix_int64_gpu('gpu_int64'),
custom_steps.compile_unix_full_gpu_no_tvm_op('gpu_no_tvm_op'),
custom_steps.compile_unix_cmake_gpu_no_tvm_op('cmake_gpu_no_tvm_op'),
custom_steps.compile_unix_cmake_gpu_no_rtc('gpu_no_rtc'),
custom_steps.compile_unix_full_gpu_mkldnn_cpp_test('gpu_mkldnn_cpp_test_make')
])
Expand All @@ -60,6 +62,8 @@ core_logic: {
// TODO(szha): fix and reenable the hanging issue. tracked in #18098
// custom_steps.test_unix_distributed_kvstore_gpu('gpu'),
custom_steps.test_unix_byteps_gpu('gpu'),
custom_steps.test_unix_python3_gpu_no_tvm_op(),
custom_steps.test_unix_capi_cpp_package('gpu_mkldnn_cpp_test_make'),
])
}
,
Expand Down
2 changes: 2 additions & 0 deletions contrib/tvmop/compile.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -152,6 +152,8 @@ def get_cuda_arch(arch):
# we create libtvmop.o first, which gives us chance to link tvm_runtime together with the libtvmop
# to allow mxnet find external helper functions in libtvm_runtime
func_binary.save(arguments.target_path + "/libtvmop.o")
if len(func_binary.imported_modules):
func_binary.imported_modules[0].save(arguments.target_path + "/libtvmop.cubin")
ld_path = arguments.target_path if arguments.ld_path is None else arguments.ld_path
create_shared(arguments.target_path + "/libtvmop.so",
arguments.target_path + "/libtvmop.o",
Expand Down
10 changes: 9 additions & 1 deletion src/c_api/c_api.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1363,7 +1363,15 @@ int MXGetVersion(int *out) {
#if MXNET_USE_TVM_OP
int MXLoadTVMOp(const char *libpath) {
API_BEGIN();
tvm::runtime::TVMOpModule::Get()->Load(libpath);
tvm::runtime::TVMOpModule *global_module = tvm::runtime::TVMOpModule::Get();
global_module->Load(libpath);
#if MXNET_USE_CUDA
std::string libpathstr(libpath);
std::string cubinpath = libpathstr.substr(0, libpathstr.size() - 11) + "libtvmop.cubin";
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

would be better to pass libpath as dir, and do libpath + "libtvmop.so" as well to keep consistency.

Copy link
Contributor Author

@jinboci jinboci Jun 12, 2020 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I rewrite it in a more elegant way :)

Yes, but MXLoadTVMOp is called at:
https://github.com/apache/incubator-mxnet/blob/1bf881f381f91b157a26d9beddcaa8f4960cc038/python/mxnet/tvmop.py#L31-L32
where _LIB_TVM_OP is returned from the
https://github.com/apache/incubator-mxnet/blob/1bf881f381f91b157a26d9beddcaa8f4960cc038/python/mxnet/libinfo.py#L25
, and _LIB_TVM_OP[0] is the path of libtvmop.so.
We may need to modify find_lib_path or write a new function to get the directory that libtvmop.so locates.

tvm::runtime::TVMOpModule cubin_module;
cubin_module.Load(cubinpath);
global_module->Import(cubin_module);
#endif
API_END();
}

Expand Down
6 changes: 6 additions & 0 deletions src/operator/tvmop/op_module.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -46,6 +46,12 @@ void TVMOpModule::Load(const std::string &filepath) {
*module_ptr_ = module;
}

void TVMOpModule::Import(const TVMOpModule& module) {
CHECK(module_ptr_ != nullptr) << "module_ptr_ is not initialized.";
std::lock_guard<std::mutex> lock(mutex_);
module_ptr_->Import(*(module.module_ptr_));
}

PackedFunc GetFunction(const std::shared_ptr<Module> &module,
const std::string &op_name,
const std::vector<mxnet::TBlob> &args) {
Expand Down
2 changes: 2 additions & 0 deletions src/operator/tvmop/op_module.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -44,6 +44,8 @@ class TVMOpModule {
// Load TVM operators binary
void Load(const std::string& filepath);

void Import(const TVMOpModule& module);

void Call(const std::string& func_name,
const mxnet::OpContext& ctx,
const std::vector<mxnet::TBlob>& args) const;
Expand Down