Weixing Zhang b4b1c6440a Enable ORT with CUDA 11 toolkit (#4168) ... * ORT on CUDA 11 1. Seperate HOROVOD and MPI 2. Seperate NCCL from HOROVOD in CMakeLists.txt 2. Remove dependency on external cub 3. cudnnSetRNNDescriptor is changed in cuDNN 8.0 * polish the code about MPI/NCCL in CMakeLists.txt and build.py * check CUDA version * ${MPI_INCLUDE_DIRS} should be PUBLIC * sm30, sm50 are deprecated in CUDA 11 Toolkit * update change based on code review feedback. * add sm_52 * improve MPI/NCCL build path Co-authored-by: Weixing Zhang <wezhan@microsoft.com>		2020-06-15 08:47:03 -07:00
..
codegen
common	Fix Windows Inbox build failing on 1) building raw api tests and 2) referencing _winml namespace in onnxruntime.dll (#3872)	2020-05-08 15:59:16 -07:00
dll
framework	fix optional input/outputs (#4229)	2020-06-15 08:10:22 +10:00
graph	Enable static memory planning for pipeline. (#4204)	2020-06-12 21:43:50 -07:00
language_interop_ops
mlas	MLAS: fuse float output into quantized GEMM (#4215)	2020-06-12 17:50:40 -07:00
optimizer	Fix subgraph based reshape fusion (#4185)	2020-06-14 21:10:08 -07:00
platform	use LOAD_WITH_ALTERED_SEARCH_PATH for LoadLibraryExA (#3908)	2020-05-11 19:53:34 -07:00
profile	Use OrtMutex and OrtCondVar everywhere instead of std::mutex/std::condition_variable for consistency.	2020-06-03 08:42:16 -07:00
protobuf	Address PR comments and clean up. (#3536)	2020-04-15 15:51:52 -07:00
providers	Enable ORT with CUDA 11 toolkit (#4168)	2020-06-15 08:47:03 -07:00
session	Fix crash reported in #4070. (#4091)	2020-06-01 15:27:14 -07:00
util	MLAS: fuse float output into quantized GEMM (#4215)	2020-06-12 17:50:40 -07:00