Tang, Cheng 8f34c8c8ed Introduce collective ops to ort inference build (#14399) ... ### Description Introduce collective ops into onnxruntime inference build, including 1) AllReduce and AllGather schema in contrib op, controlled by USE_MPI flag 2) AllReduce and AllGather kernel in cuda EP, controlled by ORT_USE_NCCL flag ### Motivation and Context Enable the collective ops in onnxruntime inference build so we have the ability to run distributed inference with multiple GPUs. The original ncclAllReduce ops in training build require quite complex configurations, which is not suitable for inference case, and it already broken. so we introduce a new implementation. --------- Co-authored-by: Cheng Tang <chenta@microsoft.com@orttrainingdev9.d32nl1ml4oruzj4qz3bqlggovf.px.internal.cloudapp.net>		2023-02-07 13:47:48 -08:00
..
android	Add onnxruntime_BUILD_UNIT_TESTS=OFF definition to iOS package build options. (#13238)	2022-10-10 18:00:17 -07:00
apple	Remove SafeInt dependency from Objective-C API. (#13698)	2022-11-18 17:06:12 -08:00
azure-pipelines	Introduce collective ops to ort inference build (#14399)	2023-02-07 13:47:48 -08:00
js	Use full ORT package for onnxruntime-react-native. (#13037)	2022-09-23 07:20:03 +10:00
linux	Add support for python 3.10 for onnxruntime-training cuda and cpu (#14100)	2023-02-02 11:32:41 -08:00
pai	Enable ccache for HIP objects (#14465)	2023-01-28 22:34:24 +08:00
python_checks
windows	upgrade protobuf to 3.20.2 and onnx to 1.13 (#14279)	2023-01-31 12:55:09 -08:00
Doxyfile_csharp.cfg