ClearML/clearml-docs
1
0
Fork 0
mirror of https://github.com/clearml/clearml-docs synced 2025-04-26 01:09:38 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

Remove GPU Compute (#823)

Browse Source
This commit is contained in:
pollfly 2024-04-11 09:59:31 +03:00 committed by GitHub
parent 1ed1a210bd
commit 73dbfe2df7
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
6 changed files with 5 additions and 100 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
docs/cloud_autoscaling/autoscaling_overview.md
View File

@ -28,7 +28,6 @@ and have the apps automatically manage your resource consumption as needed–wit
## Autoscaler Applications
ClearML provides the following GUI autoscaler applications:
* [GPU Compute](../webapp/applications/apps_gpu_compute.md) (powered by Genesis Cloud)
* [AWS Autoscaler](../webapp/applications/apps_aws_autoscaler.md)
* [GCP Autoscaler](../webapp/applications/apps_gcp_autoscaler.md)
@ -38,8 +37,6 @@ pay only for the time that you actually use the machines.
The **AWS** and **GCP** autoscaler applications will manage instances on your behalf in your cloud account. When
launching an app instance, you will provide your cloud service credentials so the autoscaler can access your account.
The **GPU Compute** application provides on-demand GPU instances powered by Genesis. All you need to do is define your compute resource budget, and you’re good to go.
## How ClearML Autoscaler Apps Work
![Autoscaler diagram](../img/autoscaler_single_queue_diagram.png)
@ -67,10 +64,7 @@ associated with a different resource, CPU and GPU, and each queue has two enqueu
the autoscaler spins up four machines, two CPU machines to execute the tasks in the CPU queue and two GPU machines to
execute the tasks in the GPU queue.
:::note
The GPU Compute app spins up a single compute resource, so you can launch multiple app instances in order to work with
multiple resources.
:::
### Task Execution Configuration
@ -90,7 +84,6 @@ executed tasks will have access to your storage service.
#### Additional Configuration
Go to a specific app’s documentation page to view all configuration options
* [GPU Compute](../webapp/applications/apps_gpu_compute.md)
* [AWS Autoscaler](../webapp/applications/apps_aws_autoscaler.md)
* [GCP Autoscaler](../webapp/applications/apps_gcp_autoscaler.md)

4
docs/getting_started/main.md
View File

@ -112,8 +112,8 @@ Want a more in depth introduction to ClearML? Choose where you want to get start
alert you whenever your model improves in accuracy)
- Manage your [data](../clearml_data/clearml_data.md) - store, track, and version control
- Remotely execute experiments on any compute resource you have available with [ClearML Agent](../clearml_agent.md)
- Automatically scale cloud instances according to your resource needs with ClearML's [GPU Compute](../webapp/applications/apps_gpu_compute.md),
[AWS Autoscaler](../webapp/applications/apps_aws_autoscaler.md), and [GCP Autoscaler](../webapp/applications/apps_gcp_autoscaler.md)
- Automatically scale cloud instances according to your resource needs with ClearML's
[AWS Autoscaler](../webapp/applications/apps_aws_autoscaler.md) and [GCP Autoscaler](../webapp/applications/apps_gcp_autoscaler.md)
GUI applications
- Run [hyperparameter optimization](../fundamentals/hpo.md)
- Build [pipelines](../pipelines/pipelines.md) from code

83
docs/webapp/applications/apps_gpu_compute.md
View File

@ -1,83 +0,0 @@
---
title: GPU Compute
---
:::info Pro Plan Offering
The ClearML GPU Compute App is available under the ClearML Pro plan
:::
Set up to run your workloads on 100% green cloud machines at optimized costs – no setup required! The ClearML GPU Compute
Application automatically spins cloud machines up or down based on demand. The app optimizes machine usage according to
a user defined resource budget: define your budget by specifying the GPU type and number of GPUs you want to use.
Each application instance monitors a ClearML queue: new cloud machines are spun up if there are pending jobs on the
queue. The app instance automatically terminates idle machines based on a specified maximum idle time.
For more information about how autoscalers work, see [Autoscalers Overview](../../cloud_autoscaling/autoscaling_overview.md#autoscaler-applications).
## GPU Compute Instance Configuration
* **Import Configuration** - Import an app instance configuration file. This will fill the configuration wizard with the
values from the file, which can be modified before launching the app instance
* **Machine Specification**
* GPU Type - NVIDIA GPU on the machine
* Number of GPUs - Number of GPUs in the cloud machine
* The rest of the machine's available resources are dependent on the number and type of GPUs specified above:
* vCPUs - Number of vCPUs in the cloud machine
* Memory - RAM available to the cloud machine
* Hourly Price - Machine's hourly rate
* Disk - Amount of Disk space available to the cloud machine
* Monitored Queue - Queue associated with application instance. The tasks enqueued to this queue will be executed on
machines of this specification
* Cloud Machine Limit - Maximum number of concurrent machines to launch
* **Idle Time Limit** (optional) - Maximum time in minutes that a cloud machine can be idle before it is spun down
* **Default Docker Image** - Default Docker image in which the ClearML Agent will run. Provide a Docker stored
in a Docker artifactory so instances can automatically fetch it
* **Git Configuration** - Git credentials with which the ClearML Agents running on your cloud instances will access your repositories to retrieve the code for their jobs
* Git User
* Git Password / Personal Access Token
* **Cloud Storage Access** (optional) - Access credentials to cloud storage service. Provides ClearML Tasks running on cloud
machines access to your storage
* **Additional ClearML Configuration** (optional) - A ClearML configuration file to use by the ClearML Agent when executing your experiments
![GPU Compute wizard](../../img/apps_gpu_compute_wizard.png)
## Dashboard
Once a GPU Compute instance is launched, the dashboard displays a summary of your cloud usage and costs.
![GPU Compute dashboard](../../img/apps_gpu_compute_dashboard.png)
The GPU Compute dashboard shows:
* Service status indicator
* <img src="/docs/latest/icons/ico-server-ok.svg" alt="Working server" className="icon size-md space-sm" /> - Cloud service is available
* <img src="/docs/latest/icons/ico-server-alert.svg" alt="Not working server" className="icon size-md space-sm" /> - Cloud service is currently unavailable
* Cloud instance details
* GPU type
* Number of GPUs
* Number of vCPUs
* RAM
* Storage
* Cost details
* Instance rate
* Total cost for current billing cycle
* Number of current running cloud instances
* Instance History - Number of running cloud instances over time
* Console - The log shows updates of cloud instances being spun up/down.
:::tip Console Debugging
To make the autoscaler console log show additional debug information, change an active app instance's log level to DEBUG:
1. Go to the app instance task's page > **CONFIGURATION** tab > **USER PROPERTIES** section
1. Hover over the section > Click `Edit` > Click `+ADD PARAMETER`
1. Input `log_level` as the key and `DEBUG` as the value of the new parameter.
![Autoscaler debugging](../../img/webapp_autoscaler_debug_log.png)
The console's log level will update in the autoscaler's next iteration.
:::
:::tip EMBEDDING CLEARML VISUALIZATION
You can embed plots from the app instance dashboard into [ClearML Reports](../webapp_reports.md). These visualizations
are updated live as the app instance(s) updates. The Enterprise Plan and Hosted Service support embedding resources in
external tools (e.g. Notion). Hover over the plot and click <img src="/docs/latest/icons/ico-plotly-embed-code.svg" alt="Embed code" className="icon size-md space-sm" />
to copy the embed code, and navigate to a report to paste the embed code.
:::

2
docs/webapp/applications/apps_overview.md
View File

@ -24,7 +24,7 @@ ClearML provides the following applications:
* [**Streamlit Launcher**](apps_streamlit.md) - Create visual web interfaces for your models with Streamlit (available under ClearML Enterprise Plan)
:::info Autoscalers
Autoscaling ([GPU Compute](apps_gpu_compute.md), [AWS Autoscaler](apps_aws_autoscaler.md), and [GCP Autoscaler](apps_gcp_autoscaler.md))
Autoscaling ([AWS Autoscaler](apps_aws_autoscaler.md) and [GCP Autoscaler](apps_gcp_autoscaler.md))
was previously available through the Applications page. The autoscaler functionality has been moved to the [Orchestration page](https://app.clear.ml/workers-and-queues/autoscalers)
in the WebApp.
:::

6
docs/webapp/webapp_workers_queues.md
View File

@ -24,7 +24,6 @@ The ClearML Autoscaler apps are available under the ClearML Pro plan
Use the **AUTOSCALERS** tab to access ClearML's cloud autoscaling applications:
* GPU Compute (powered by Genesis Cloud)
* AWS Autoscaler
* GCP Autoscaler
@ -34,14 +33,11 @@ pay only for the time that you actually use the machines.
The **AWS** and **GCP** autoscaler applications will manage instances on your behalf in your cloud account. When
launching an app instance, you will provide your cloud service credentials so the autoscaler can access your account.
The **GPU Compute** application provides on-demand GPU instances powered by Genesis. All you need to do is define your
compute resource budget, and you're good to go.
Once you launch an autoscaler app instance, you can monitor the autoscaler's activity and your cloud usage in the instance's
dashboard.
For more information about how autoscalers work, see the [Cloud Autoscaling Overview](../cloud_autoscaling/autoscaling_overview.md).
For more information about a specific autoscaler, see [GPU Compute](applications/apps_gpu_compute.md), [AWS Autoscaler](applications/apps_aws_autoscaler.md),
For more information about a specific autoscaler, see [AWS Autoscaler](applications/apps_aws_autoscaler.md)
and/or [GCP Autoscaler](applications/apps_gcp_autoscaler.md).
![Cloud autoscalers](../img/webapp_orchestration_autoscalers.png)

1
sidebars.js
View File

@ -40,7 +40,6 @@ module.exports = {
{'Cloud Autoscaling': [
'cloud_autoscaling/autoscaling_overview',
{'Autoscaler Apps': [
'webapp/applications/apps_gpu_compute',
'webapp/applications/apps_aws_autoscaler',
'webapp/applications/apps_gcp_autoscaler',
]