PR #4168 Enable ORT with CUDA 11 toolkit

Enable ORT with CUDA 11 toolkit #4168

weixingzhang merged 8 commits into master from wezhan/cuda11

weixingzhang added training

weixingzhang requested a review from

snnn 5 years ago

weixingzhang requested a review from

SherlockNoMad 5 years ago

weixingzhang requested a review from

ke1337 5 years ago

weixingzhang requested a review from

HectorSVC 5 years ago

weixingzhang requested a review 5 years ago

weixingzhang added core runtime

Tixxx commented on 2020-06-09

HectorSVC commented on 2020-06-09

snnn commented on 2020-06-09

ORT on CUDA 11

ddb2a12f

polish the code about MPI/NCCL in CMakeLists.txt and build.py

95891965

check CUDA version

2d1a9d01

${MPI_INCLUDE_DIRS} should be PUBLIC

0a97ad12

sm30, sm50 are deprecated in CUDA 11 Toolkit

8a603310

update change based on code review feedback.

f005c1f8

add sm_52

9d7d2549

improve MPI/NCCL build path

fe6b0816

weixingzhang force pushed from 58046606 to fe6b0816 5 years ago

snnn approved these changes on 2020-06-15

Tixxx approved these changes on 2020-06-15

weixingzhang merged b4b1c644 into master 5 years ago

weixingzhang deleted the wezhan/cuda11 branch 5 years ago

Reviewers

Tixxx

snnn

HectorSVC

SherlockNoMad

ke1337

Assignees

No one assigned

Labels

training core runtime

Milestone

No milestone

onnxruntime Enable ORT with CUDA 11 toolkit #4168 Merged

Enable ORT with CUDA 11 toolkit #4168

onnxruntime
Enable ORT with CUDA 11 toolkit
#4168

Merged