Added GPU-Access to systemd-Service Definition (#10694)

* Update parent title for install and configure benchmark pages (#10686)

Signed-off-by: Fanit Kolchina <[email protected]>
Signed-off-by: Konstantin Nebel <[email protected]>

* Delete unused images from the site and correct image formatting (#10691)

* Remove unused images

Signed-off-by: Fanit Kolchina <[email protected]>

* Delete icons file and correct image formatting

Signed-off-by: Fanit Kolchina <[email protected]>

* Heading formatting

Signed-off-by: Fanit Kolchina <[email protected]>

---------

Signed-off-by: Fanit Kolchina <[email protected]>
Signed-off-by: Konstantin Nebel <[email protected]>

* Added GPU-Access for systemd

---------

Signed-off-by: Konstantin Nebel <[email protected]>

* Update _ml-commons-plugin/gpu-acceleration.md

Co-authored-by: kolchfa-aws <[email protected]>
Signed-off-by: TheBigK02 <[email protected]>

* Update _ml-commons-plugin/gpu-acceleration.md

Co-authored-by: kolchfa-aws <[email protected]>
Signed-off-by: TheBigK02 <[email protected]>

---------

Signed-off-by: Fanit Kolchina <[email protected]>
Signed-off-by: Konstantin Nebel <[email protected]>
Signed-off-by: TheBigK02 <[email protected]>
Co-authored-by: kolchfa-aws <[email protected]>
Co-authored-by: Konstantin Nebel <[email protected]>
(cherry picked from commit aa12026)
Signed-off-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>

Author: 3 people

_ml-commons-plugin/gpu-acceleration.md

@@ -67,6 +67,18 @@ else
 fi
 ```
 
+If you run OpenSearch natively (without Docker) using the packaged version of OpenSearch, `systemd` may block OpenSearch from accessing your GPU. To accelerate models, you need a working [CUDA Toolkit](https://developer.nvidia.com/cuda-toolkit) installation and access to the NVIDIA device under `/dev`.
+
+To allow OpenSearch to use the GPU, update the `systemd` service by adding the following configuration:
+
+```ini
+systemctl edit opensearch.service
+
+[Service]
+DevicePolicy=auto
+``` 
+
+
 After verifying that `nvidia-uvm` exists under `/dev`, you can start OpenSearch inside your cluster. 
 
 ### Preparing AWS Inferentia ML node