ClearML/clearml-docs
1
0
Fork 0
mirror of https://github.com/clearml/clearml-docs synced 2025-05-06 22:05:31 +00:00
Code Issues Actions 1 Packages Projects Releases Wiki Activity

Fix clearml-agent dynamic gpu example (#43)

Browse Source 
add --gpus flag
fix --queue flag
This commit is contained in:
pollfly 2021-08-22 15:08:57 +03:00 committed by GitHub
parent d5942a025c
commit df935fa049
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
2 changed files with 3 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
docs/clearml_agent.md
View File

@ -487,7 +487,7 @@ An agent can be spun on multiple GPUs (e.g. 8 GPUs, `--gpus 0-7`), and then atta
queues that are configured to run with a certain amount of resources: queues that are configured to run with a certain amount of resources:
```console ```console
clearml-agent daemon --dynamic-gpus --queues quad_gpu=4 dual_gpu=2 clearml-agent daemon --dynamic-gpus --gpus 0-7 --queue quad_gpu=4 dual_gpu=2
``` ```
The agent can now spin multiple Tasks from the different queues based on the number of GPUs configured to the queue. The agent can now spin multiple Tasks from the different queues based on the number of GPUs configured to the queue.
@ -497,7 +497,7 @@ queue, look for available GPUs again and spin on GPUs 4-5.
Another option for allocating GPUs: Another option for allocating GPUs:
```console ```console
clearml-agent daemon --dynamic-gpus --queue dual=2 opportunistic=1-4 clearml-agent daemon --dynamic-gpus --gpus 0-7 --queue dual=2 opportunistic=1-4
``` ```
Notice that a minimum and maximum value of GPUs was specified for the `opportunistic` queue. This means the agent Notice that a minimum and maximum value of GPUs was specified for the `opportunistic` queue. This means the agent

2
docs/references/clearml_agent_ref.md
View File

@ -213,7 +213,7 @@ Use the `daemon` subcommand to run a worker, optionally in a Docker container, l
**`dynamic-gpus`** **`dynamic-gpus`**
* Allow to dynamically allocate gpus based on queue properties, configure with `--queues <queue_name>=<num_gpus>`. * Allow to dynamically allocate gpus based on queue properties, configure with `--queue <queue_name>=<num_gpus>`.
For example: `--dynamic-gpus --queue dual_gpus=2 single_gpu=1` For example: `--dynamic-gpus --queue dual_gpus=2 single_gpu=1`