Compile Error

[zhongyuansh]

env: Ubuntu 20.04 GCC-10

error: torch-ccl/third_party/oneCCL/src/atl/util/pm/pmi_resizable_rt/pmi_resizable_simple.h:124:17: error: field ‘my_proccess_name’ has incomplete type ‘std::string’ {aka ‘std::__cxx11::basic_string’}
124 | std::string my_proccess_name;

torch-ccl/third_party/oneCCL/src/atl/util/pm/pmi_resizable_rt/pmi_resizable_simple.cpp:50:36: error: ‘getenv’ was not declared in this scope

why?

[chengjunlu]

@zhongyuansh It seems a c++ library compatible issue.

Which version of the torch_ccl are you using?

[zhongyuansh]

I am using the latest version (master branch)

I have fixed this issue add "#include < string >" in pmi_resizable_simple.h file, these compilication errors will be fixed

But still hava another issue, as following: torch-ccl/third_party/oneCCL/src/comp/bf16/bf16_intrisics.hpp:74:82: note: use ‘-flax-vector-conversions’ to permit conversions between vectors with differing element types or numbers of subparts
74 | _mm256_storeu_si256((__m256i*)(dst), _mm512_cvtneps_pbh(_mm512_loadu_ps(src)));
| t torch-ccl/third_party/oneCCL/src/comp/bf16/bf16_intrisics.hpp:74:60: error: cannot convert ‘__m256bh’ to ‘__m256i’
74 | _mm256_storeu_si256((__m256i*)(dst), _mm512_cvtneps_pbh(_mm512_loadu_ps(src)));
| ~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~
| |
| __m256bh

how to fix this issue? where should the "-flax-vector-conversions" flag be added?

[chengjunlu]

@zhongyuansh I have create a oneCCL issue about the compile error in oneCCL.

Could you upload the full error long of building the torch_ccl?