5 Commits (8894c67d7168dfa5f5dc8e57ec2b5c60f24e368c)

Author	SHA1	Message	Date
QI JUN	61ec0b9516	Refine device context (#6433) ... There are mainly following fixes: - take `DeviceContext` as the template parameter of math functors and OpKernel instead of `Place` - remove `eigen_device` interface in base class `DeviceContext` - remove `GetEigenDevice` interface in `ExecutionContext` and base class `DeviceContext` - remove unused `platform::EigenDeviceConverter` - rename `REGISTER_OP_GPU_KERNEL` to `REGISTER_OP_CUDA_KERNEL` - rename `USE_GPU_ONLY_OP` to `USE_CUDA_ONLY_OP`	7 years ago
chengduoZH	bc3ec53671	remove conflict	7 years ago
chengduoZH	5238c9fb0b	input type should be different	7 years ago
chengduoZH	3bd3cc0c85	add double type for kernel	7 years ago
dangqingqing	f5e367655e	Use G++ to compile some cu operators.	7 years ago