Error compile for GPU support
- Dominant language
- C++
- Stars
- 3.4k
- Forks
- 701
- PR merge metrics
- No merged PRs in 30d
Description
gcc version: 4.8.5
nvcc version: 5.5
It seems that these code are generated. Hope you can help me
```
/home/zhuyi/repos/eigen/unsupported/Eigen/CXX11/src/Tensor/TensorEvaluator.h(132): error: calling a constexpr __host__ function("std::round") from a __device__ function("Eigen::internal::EigenMetaKernelEval< ::Eigen::TensorEvaluator , (int)0, ::Eigen::MakePointer> , const ::Eigen::TensorCwiseUnaryOp< ::Eigen::internal::scalar_round_op , const ::Eigen::TensorMap< ::Eigen::Tensor , (int)0, ::Eigen::MakePointer> > > , ::Eigen::GpuDevice> , long, (bool)0> ::run") is not allowed. The experimental flag '--expt-relaxed-constexpr' can be used to allow this.
/home/zhuyi/repos/eigen/unsupported/Eigen/CXX11/src/Tensor/TensorEvaluator.h(132): error: identifier "std::round" is undefined in device code
```
Contributor guide
No contributing guide indexed for this repository
Research direction
Start by reproducing the GPU build with gcc 4.8.5 and nvcc 5.5, then inspect unsupported/Eigen/CXX11/src/Tensor/TensorEvaluator.h around line 132. Compare the compiler diagnostics for std::round in device code and determine whether the reported configuration can compile successfully; the issue is done when the GPU-supported build completes without these errors.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- cpp
- Domain
- build-system
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100