Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Remove memory copy between TensorRT and CUDA #1561

Merged
merged 23 commits into from
Aug 9, 2019
Merged

Remove memory copy between TensorRT and CUDA #1561

Changes from 1 commit
Commits
Show all changes
23 commits
Select commit Hold shift + click to select a range
35c9e96
remove memory copy between CUDA and TRT
stevenlix Aug 1, 2019
c66fbe8
add info to RegisterExecutionProvider input
stevenlix Aug 1, 2019
bb6e268
use new IDeviceAllocator for trt allocator
stevenlix Aug 1, 2019
3cad0ba
remove SetDefaultInputsMemoryType from TRT EP
stevenlix Aug 1, 2019
a5d1893
Merge branch 'master' into stevenlix/trtoverhead
stevenlix Aug 1, 2019
231cf0b
remove onnx-tensorrt 5.0
stevenlix Aug 1, 2019
a7f534e
add submodule onnx-tensorrt branch 5.1
stevenlix Aug 2, 2019
4d7812f
remove redundancy
stevenlix Aug 5, 2019
7aab699
Update transformer_memcpy.cc
stevenlix Aug 5, 2019
7f306ff
Update tensorrt_execution_provider.cc
stevenlix Aug 5, 2019
5cc3e1c
switch to TensorRT 5.1.5.0
stevenlix Aug 5, 2019
7cbd863
Merge branch 'stevenlix/trtoverhead' of https://github.com/Microsoft/…
stevenlix Aug 5, 2019
196386a
update python binding
stevenlix Aug 6, 2019
5c2d5e5
disable failed test case on TensorRT
stevenlix Aug 6, 2019
bd604e9
Merge branch 'master' into stevenlix/trtoverhead
stevenlix Aug 6, 2019
96c7560
Update activation_op_test.cc
stevenlix Aug 6, 2019
62ec7cc
upgrade to TensorRT container 19.06
stevenlix Aug 6, 2019
40240d0
Merge branch 'stevenlix/trtoverhead' of https://github.com/Microsoft/…
stevenlix Aug 6, 2019
3bee98d
update according to feedback
stevenlix Aug 8, 2019
94923d7
add comments
stevenlix Aug 8, 2019
5d720b0
remove tensorrt allocator and use cuda(gpu) allocator
stevenlix Aug 8, 2019
b9cb8f1
update onnx-tensorrt submodule
stevenlix Aug 8, 2019
f153d91
change ci build cuda directory name
stevenlix Aug 8, 2019
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Prev Previous commit
Next Next commit
upgrade to TensorRT container 19.06
  • Loading branch information
@stevenlix
stevenlix committed Aug 6, 2019
commit 62ec7cca23007e2ee5a1be435c7867508de48b66
8 changes: 4 additions & 4 deletions tools/ci_build/github/linux/docker/Dockerfile.ubuntu_tensorrt
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,8 +1,8 @@
# Tag: nvcr.io/nvidia/tensorrt:19.02-py3
# Label: com.nvidia.cuda.version: 10.0.130
# Label: com.nvidia.cudnn.version: 7.4.2
# Tag: nvcr.io/nvidia/tensorrt:19.06-py3
# Label: com.nvidia.cuda.version: 10.1.168
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

so are we using different cuda/cudnn versions for windows vs linux?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Linux CI machine is using docker container, where the cuda/cudnn version is chosen by Nvidia. Ideally windows machine should also update cuda/cudnn to their new versions if there is no side effect.

# Label: com.nvidia.cudnn.version: 7.6.0
# Ubuntu 16.04
FROM nvcr.io/nvidia/tensorrt:19.02-py3
FROM nvcr.io/nvidia/tensorrt:19.06-py3

ARG PYTHON_VERSION=3.5

Expand Down