Skip to content

Commit

Permalink
Merge branch 'main' into chhwang/cast
Browse files Browse the repository at this point in the history
  • Loading branch information
chhwang authored Sep 25, 2023
2 parents a97a562 + 8e4a723 commit f383ddf
Show file tree
Hide file tree
Showing 2 changed files with 20 additions and 9 deletions.
27 changes: 19 additions & 8 deletions docs/install.md
Original file line number Diff line number Diff line change
Expand Up @@ -16,11 +16,6 @@

- Compute capability 9.0 support will be added in the future.

* To run ARK in a Docker container, we need to mount `/dev` and `/lib/modules` into the container so that the container can use `gpumem` driver. Add the following options in the `docker run` command:
```
-v /dev:/dev -v /lib/modules:/lib/modules
```
* Mellanox OFED

## Docker Images
Expand All @@ -34,9 +29,25 @@ docker pull ghcr.io/microsoft/ark/ark:base-cuda12.1

Check [ARK containers](https://github.com/microsoft/ark/pkgs/container/ark%2Fark) for all available Docker images.

To run ARK in a Docker container, we need to mount `/dev` and `/lib/modules` into the container so that the container can use `gpumem` driver. Specifically, add `--privileged -v /dev:/dev -v /lib/modules:/lib/modules` in the `docker run` command. The following is an example.
```
docker run \
--privileged \
--cap-add=ALL \
--shm-size=1g \
--ulimit memlock=-1 \
--ulimit stack=67108864 \
--net=host \
--ipc=host \
--gpus all \
-v /dev:/dev \
-v /lib/modules:/lib/modules \
-it --name [Container Name] [Image Name] bash
```

## Install `gpudma`

*NOTE: if you are using a Docker container, the following steps should be done on the host.*
**NOTE: if you are using a Docker container, the steps in this section should be done on the host.**

1. Pull submodules.

Expand All @@ -55,7 +66,7 @@ Check [ARK containers](https://github.com/microsoft/ark/pkgs/container/ark%2Fark
3. Load `gpumem` driver.
```bash
sudo insmod third_party/gpudma/module/gpumem.ko
sudo insmod gpudma/module/gpumem.ko
sudo chmod 666 /dev/gpumem
```
Expand Down Expand Up @@ -83,7 +94,7 @@ Check [ARK containers](https://github.com/microsoft/ark/pkgs/container/ark%2Fark
```bash
cd examples/tutorial
python3 tutorial.py
python3 quickstart_tutorial.py
```
## (Optional) Install ARK C++ and Run Unit Tests
Expand Down
2 changes: 1 addition & 1 deletion docs/quickstart.md
Original file line number Diff line number Diff line change
Expand Up @@ -8,7 +8,7 @@ Please refer to the [ARK Install Instructions](./install.md) to install ARK for

## Quick Start Tutorial

You can run a tutorial example at [tutorial.py](../examples/tutorial/quickstart_tutorial.py) to see how ARK works.
You can run a tutorial example at [quickstart_tutorial.py](../examples/tutorial/quickstart_tutorial.py) to see how ARK works.

```bash
python examples/tutorial/quickstart_tutorial.py
Expand Down

0 comments on commit f383ddf

Please sign in to comment.