[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#1064484: pytorch-cuda: FTBFS: error: cannot convert ‘dnnl::memory::data_type’ to ‘const ideep::dims&’



Source: pytorch-cuda
Version: 2.0.1+dfsg-5
Severity: serious
Tags: ftbfs
Justification: fails to build from source

Hi,

while locally rebuilding all packages build-depending on
nvidia-cuda-toolkit, I noticed that pytorch-cuda started to FTBFS:

[621/2435] /usr/bin/cuda-g++ -DAT_PER_OPERATOR_HEADERS -DFMT_HEADER_ONLY=1 -DHAVE_MALLOC_USABLE_SIZE=1 -DHAVE_MMAP=1 -DHAVE_SHM_OPEN=1 -DHAVE_SHM_UNLINK=1 -DMINIZ_DISABLE_ZIP_READER_CRC32_CHECKS -DONNXIFI_ENABLE_EXT=1 -DONNX_ML=1 -DONNX_NAMESPACE=onnx -DUSE_
C10D_GLOO -DUSE_DISTRIBUTED -DUSE_EXTERNAL_MZCRC -DUSE_FLASH_ATTENTION -DUSE_RPC -DUSE_TENSORPIPE -D_FILE_OFFSET_BITS=64 -Dtorch_cpu_EXPORTS -I/build/pytorch-cuda-2.0.1+dfsg/build/aten/src -I/build/pytorch-cuda-2.0.1+dfsg/aten/src -I/build/pytorch-cuda-2.0.1
+dfsg/build -I/build/pytorch-cuda-2.0.1+dfsg -I/build/pytorch-cuda-2.0.1+dfsg/cmake/../third_party/benchmark/include -I/build/pytorch-cuda-2.0.1+dfsg/debian/foxi -I/build/pytorch-cuda-2.0.1+dfsg/build/debian/foxi -I/build/pytorch-cuda-2.0.1+dfsg/torch/csrc/a
pi -I/build/pytorch-cuda-2.0.1+dfsg/torch/csrc/api/include -I/build/pytorch-cuda-2.0.1+dfsg/caffe2/aten/src/TH -I/build/pytorch-cuda-2.0.1+dfsg/build/caffe2/aten/src/TH -I/build/pytorch-cuda-2.0.1+dfsg/build/caffe2/aten/src -I/build/pytorch-cuda-2.0.1+dfsg/b
uild/caffe2/../aten/src -I/build/pytorch-cuda-2.0.1+dfsg/torch/csrc -I/build/pytorch-cuda-2.0.1+dfsg/third_party/miniz-2.1.0 -I/build/pytorch-cuda-2.0.1+dfsg/debian/kineto/libkineto/include -I/build/pytorch-cuda-2.0.1+dfsg/debian/kineto/libkineto/src -I/buil
d/pytorch-cuda-2.0.1+dfsg/aten/../third_party/catch/single_include -I/build/pytorch-cuda-2.0.1+dfsg/aten/src/ATen/.. -I/build/pytorch-cuda-2.0.1+dfsg/c10/.. -I/build/pytorch-cuda-2.0.1+dfsg/aten/src/ATen/native/quantized/cpu/qnnpack/include -I/build/pytorch-
cuda-2.0.1+dfsg/aten/src/ATen/native/quantized/cpu/qnnpack/src -I/build/pytorch-cuda-2.0.1+dfsg/aten/src/ATen/native/quantized/cpu/qnnpack/deps/clog/include -I/build/pytorch-cuda-2.0.1+dfsg/third_party/flatbuffers/include -isystem /build/pytorch-cuda-2.0.1+d
fsg/build/third_party/gloo -isystem /build/pytorch-cuda-2.0.1+dfsg/cmake/../third_party/gloo -isystem /build/pytorch-cuda-2.0.1+dfsg/cmake/../third_party/googletest/googlemock/include -isystem /build/pytorch-cuda-2.0.1+dfsg/cmake/../third_party/googletest/go
ogletest/include -isystem /usr/include/opencv4 -isystem /usr/include/eigen3 -isystem /build/pytorch-cuda-2.0.1+dfsg/caffe2 -Wdate-time -D_FORTIFY_SOURCE=2 -g -O2 -ffile-prefix-map=/build/pytorch-cuda-2.0.1+dfsg=. -fstack-protector-strong -fstack-clash-protec
tion -Wformat -Werror=format-security -fcf-protection -gsplit-dwarf -Wno-dangling-reference -fuse-ld=lld -I/usr -D_GLIBCXX_USE_CXX11_ABI=1 -Wno-deprecated -fvisibility-inlines-hidden -DUSE_PTHREADPOOL -DUSE_KINETO -DLIBKINETO_NOCUPTI -DLIBKINETO_NOROCTRACER
-DUSE_PYTORCH_QNNPACK -DUSE_XNNPACK -DSYMBOLICATE_MOBILE_DEBUG_HANDLE -O2 -fPIC -Wall -Wextra -Werror=return-type -Werror=non-virtual-dtor -Werror=range-loop-construct -Werror=bool-operation -Wnarrowing -Wno-missing-field-initializers -Wno-type-limits -Wno-a
rray-bounds -Wno-unknown-pragmas -Wunused-local-typedefs -Wno-unused-parameter -Wno-unused-function -Wno-unused-result -Wno-strict-overflow -Wno-strict-aliasing -Wno-error=deprecated-declarations -Wno-stringop-overflow -Wno-psabi -Wno-error=pedantic -Wno-err
or=redundant-decls -Wno-error=old-style-cast -fdiagnostics-color=always -faligned-new -Wno-unused-but-set-variable -Wno-maybe-uninitialized -fno-math-errno -fno-trapping-math -Werror=format -Werror=cast-function-type -Wno-stringop-overflow -DHAVE_AVX512_CPU_
DEFINITION -DHAVE_AVX2_CPU_DEFINITION -O2 -g -DNDEBUG -std=gnu++17 -fPIC -DCAFFE2_USE_GLOO -DTH_HAVE_THREAD -Wall -Wextra -Wno-unused-parameter -Wno-unused-function -Wno-unused-result -Wno-missing-field-initializers -Wno-write-strings -Wno-unknown-pragmas -W
no-type-limits -Wno-array-bounds -Wno-sign-compare -Wno-strict-overflow -Wno-strict-aliasing -Wno-error=deprecated-declarations -Wno-missing-braces -Wno-maybe-uninitialized -fvisibility=hidden -O2 -DCAFFE2_BUILD_MAIN_LIB -fopenmp -Wno-deprecated-declarations
 -MD -MT caffe2/CMakeFiles/torch_cpu.dir/__/aten/src/ATen/native/quantized/cpu/qlinear_prepack.cpp.o -MF caffe2/CMakeFiles/torch_cpu.dir/__/aten/src/ATen/native/quantized/cpu/qlinear_prepack.cpp.o.d -o caffe2/CMakeFiles/torch_cpu.dir/__/aten/src/ATen/native/
quantized/cpu/qlinear_prepack.cpp.o -c /build/pytorch-cuda-2.0.1+dfsg/aten/src/ATen/native/quantized/cpu/qlinear_prepack.cpp
FAILED: caffe2/CMakeFiles/torch_cpu.dir/__/aten/src/ATen/native/quantized/cpu/qlinear_prepack.cpp.o
/usr/bin/cuda-g++ -DAT_PER_OPERATOR_HEADERS -DFMT_HEADER_ONLY=1 -DHAVE_MALLOC_USABLE_SIZE=1 -DHAVE_MMAP=1 -DHAVE_SHM_OPEN=1 -DHAVE_SHM_UNLINK=1 -DMINIZ_DISABLE_ZIP_READER_CRC32_CHECKS -DONNXIFI_ENABLE_EXT=1 -DONNX_ML=1 -DONNX_NAMESPACE=onnx -DUSE_C10D_GLOO -
DUSE_DISTRIBUTED -DUSE_EXTERNAL_MZCRC -DUSE_FLASH_ATTENTION -DUSE_RPC -DUSE_TENSORPIPE -D_FILE_OFFSET_BITS=64 -Dtorch_cpu_EXPORTS -I/build/pytorch-cuda-2.0.1+dfsg/build/aten/src -I/build/pytorch-cuda-2.0.1+dfsg/aten/src -I/build/pytorch-cuda-2.0.1+dfsg/build
 -I/build/pytorch-cuda-2.0.1+dfsg -I/build/pytorch-cuda-2.0.1+dfsg/cmake/../third_party/benchmark/include -I/build/pytorch-cuda-2.0.1+dfsg/debian/foxi -I/build/pytorch-cuda-2.0.1+dfsg/build/debian/foxi -I/build/pytorch-cuda-2.0.1+dfsg/torch/csrc/api -I/build
/pytorch-cuda-2.0.1+dfsg/torch/csrc/api/include -I/build/pytorch-cuda-2.0.1+dfsg/caffe2/aten/src/TH -I/build/pytorch-cuda-2.0.1+dfsg/build/caffe2/aten/src/TH -I/build/pytorch-cuda-2.0.1+dfsg/build/caffe2/aten/src -I/build/pytorch-cuda-2.0.1+dfsg/build/caffe2
/../aten/src -I/build/pytorch-cuda-2.0.1+dfsg/torch/csrc -I/build/pytorch-cuda-2.0.1+dfsg/third_party/miniz-2.1.0 -I/build/pytorch-cuda-2.0.1+dfsg/debian/kineto/libkineto/include -I/build/pytorch-cuda-2.0.1+dfsg/debian/kineto/libkineto/src -I/build/pytorch-c
uda-2.0.1+dfsg/aten/../third_party/catch/single_include -I/build/pytorch-cuda-2.0.1+dfsg/aten/src/ATen/.. -I/build/pytorch-cuda-2.0.1+dfsg/c10/.. -I/build/pytorch-cuda-2.0.1+dfsg/aten/src/ATen/native/quantized/cpu/qnnpack/include -I/build/pytorch-cuda-2.0.1+
dfsg/aten/src/ATen/native/quantized/cpu/qnnpack/src -I/build/pytorch-cuda-2.0.1+dfsg/aten/src/ATen/native/quantized/cpu/qnnpack/deps/clog/include -I/build/pytorch-cuda-2.0.1+dfsg/third_party/flatbuffers/include -isystem /build/pytorch-cuda-2.0.1+dfsg/build/t
hird_party/gloo -isystem /build/pytorch-cuda-2.0.1+dfsg/cmake/../third_party/gloo -isystem /build/pytorch-cuda-2.0.1+dfsg/cmake/../third_party/googletest/googlemock/include -isystem /build/pytorch-cuda-2.0.1+dfsg/cmake/../third_party/googletest/googletest/in
clude -isystem /usr/include/opencv4 -isystem /usr/include/eigen3 -isystem /build/pytorch-cuda-2.0.1+dfsg/caffe2 -Wdate-time -D_FORTIFY_SOURCE=2 -g -O2 -ffile-prefix-map=/build/pytorch-cuda-2.0.1+dfsg=. -fstack-protector-strong -fstack-clash-protection -Wform
at -Werror=format-security -fcf-protection -gsplit-dwarf -Wno-dangling-reference -fuse-ld=lld -I/usr -D_GLIBCXX_USE_CXX11_ABI=1 -Wno-deprecated -fvisibility-inlines-hidden -DUSE_PTHREADPOOL -DUSE_KINETO -DLIBKINETO_NOCUPTI -DLIBKINETO_NOROCTRACER -DUSE_PYTOR
CH_QNNPACK -DUSE_XNNPACK -DSYMBOLICATE_MOBILE_DEBUG_HANDLE -O2 -fPIC -Wall -Wextra -Werror=return-type -Werror=non-virtual-dtor -Werror=range-loop-construct -Werror=bool-operation -Wnarrowing -Wno-missing-field-initializers -Wno-type-limits -Wno-array-bounds
 -Wno-unknown-pragmas -Wunused-local-typedefs -Wno-unused-parameter -Wno-unused-function -Wno-unused-result -Wno-strict-overflow -Wno-strict-aliasing -Wno-error=deprecated-declarations -Wno-stringop-overflow -Wno-psabi -Wno-error=pedantic -Wno-error=redundan
t-decls -Wno-error=old-style-cast -fdiagnostics-color=always -faligned-new -Wno-unused-but-set-variable -Wno-maybe-uninitialized -fno-math-errno -fno-trapping-math -Werror=format -Werror=cast-function-type -Wno-stringop-overflow -DHAVE_AVX512_CPU_DEFINITION
-DHAVE_AVX2_CPU_DEFINITION -O2 -g -DNDEBUG -std=gnu++17 -fPIC -DCAFFE2_USE_GLOO -DTH_HAVE_THREAD -Wall -Wextra -Wno-unused-parameter -Wno-unused-function -Wno-unused-result -Wno-missing-field-initializers -Wno-write-strings -Wno-unknown-pragmas -Wno-type-lim
its -Wno-array-bounds -Wno-sign-compare -Wno-strict-overflow -Wno-strict-aliasing -Wno-error=deprecated-declarations -Wno-missing-braces -Wno-maybe-uninitialized -fvisibility=hidden -O2 -DCAFFE2_BUILD_MAIN_LIB -fopenmp -Wno-deprecated-declarations -MD -MT ca
ffe2/CMakeFiles/torch_cpu.dir/__/aten/src/ATen/native/quantized/cpu/qlinear_prepack.cpp.o -MF caffe2/CMakeFiles/torch_cpu.dir/__/aten/src/ATen/native/quantized/cpu/qlinear_prepack.cpp.o.d -o caffe2/CMakeFiles/torch_cpu.dir/__/aten/src/ATen/native/quantized/c
pu/qlinear_prepack.cpp.o -c /build/pytorch-cuda-2.0.1+dfsg/aten/src/ATen/native/quantized/cpu/qlinear_prepack.cpp
/build/pytorch-cuda-2.0.1+dfsg/aten/src/ATen/native/quantized/cpu/qlinear_prepack.cpp: In static member function ‘static c10::intrusive_ptr<LinearPackedParamsBase> PackedLinearWeightsOnednn::prepack(at::Tensor, c10::optional<at::Tensor>)’:
/build/pytorch-cuda-2.0.1+dfsg/aten/src/ATen/native/quantized/cpu/qlinear_prepack.cpp:249:103: error: cannot convert ‘dnnl::memory::data_type’ to ‘const ideep::dims&’ {aka ‘const std::vector<long int>&’}
  249 |   auto w_desc = ideep::matmul_forward::expected_weights_desc(wgt.get_dims(), dnnl::memory::data_type::s8,
      |                                                                              ~~~~~~~~~~~~~~~~~~~~~~~~~^~
      |                                                                                                       |
      |                                                                                                       dnnl::memory::data_type
In file included from /usr/include/ideep/computations.hpp:20,
                 from /usr/include/ideep.hpp:41,
                 from /build/pytorch-cuda-2.0.1+dfsg/aten/src/ATen/native/quantized/cpu/OnednnUtils.h:7,
                 from /build/pytorch-cuda-2.0.1+dfsg/aten/src/ATen/native/quantized/cpu/qlinear_prepack.cpp:8:
/usr/include/ideep/operators/matmul.hpp:662:19: note:   initializing argument 2 of ‘static ideep::tensor::desc ideep::matmul_forward::expected_weights_desc(const ideep::dims&, const ideep::dims&, ideep::data_type, ideep::data_type, const ideep::attr_t&, cons
t ideep::engine&)’
  662 |       const dims& src_dims = dims(),
      |       ~~~~~~~~~~~~^~~~~~~~~~~~~~~~~
At global scope:
cc1plus: note: unrecognized command-line option ‘-Wno-dangling-reference’ may have been intended to silence earlier diagnostics


Andreas

Attachment: pytorch.sid.log.gz
Description: application/gzip


Reply to: