Scribd
Upload
17 views3 pages

Pages From Introduction To Einstein HPC Portal-V3-2

Uploaded by

anissameh2012
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
17 views3 pages

Pages From Introduction To Einstein HPC Portal-V3-2

Uploaded by

anissameh2012
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

To add a workflow to your Workflow tab, select the Add Parallel Workflow button on the

desired workflow tile. You will now see the selected application in the Workflow tab.

Getting Started

Once you have logged in for the first time, you will need to perform the following tasks to run
computations on the HPC system via Parallel Works:

1. Go to the User menu, select Marketplace, then add a Parallel Workflow.


2. Select the Workflows tab, then select the heart icon to the right of the desired workflow.
3. Select the Compute tab.
4. At the bottom of the window, inspect the Computing Resources area; the power button
under EINSTEIN_DEFAULT should be green. If it is not, then click the button and wait
for it to turn green. You will not be able to connect to the HPC system if this button is
gray.

All the information and controls you will need are now on the Compute tab. Your selected
workflow will be in the Favorite Workflows area.

To start a workflow, select the desired workflow icon from the Favorite Workflows area and the
system will automatically take you to the Workflows page, which will now show a form for you
to fill out. Each workflow will have different requirements; however, there are some general
specifications shared by all workflows:

● SLURM partition – Fill in the appropriate partition (queue) name. Each partition has
different time and resource restrictions, so make sure to put in the correct one for your
job.
● CPUs per task – Fill in the number of CPU cores you require. In many cases, you can
use the default of 1.
● Memory – Fill in the GB of memory your job will require. If the application attempts to
access more memory than specified, it will fail, often without any warning; insufficient
memory is a common reason for “mystery” job failure.
● Wall time – Fill in the maximum amount of time allowed for the job to complete
● Other Scheduler directives – Many forms allow you to add other SLURM directives for
fine tuning compute resource requests; these parameters may not be necessary for all
workflow runs
● Execute – Below the form will be a button to execute your commands and send the
information to the HPC system.

Monitoring and Controlling the Job

Once you select Execute, the platform will take a few minutes to set up the job and submit it to
the HPC system. You can navigate back to the Compute tab to see your job status in the
Workflow Monitor, which displays the following:

● ID – the job ID assigned by Parallel Works (not SLURM)


● Workflow – the name of the workflow
● Status – this will be Running for active jobs; afterwards you may see Completed,
Deleted, or Error, depending on the final status.
● Submitted – time that the job was created
● Runtime – amount of time so far or total, in the case of ended jobs
● Red circle – select this to cancel the job
● Blue circular arrow – this will run the workflow again (which will take you to the form)
● Eye – this will take you back to the Workflows screen where your job is executing. If this
is an interactive job, this will display the application GUI.

Your job is now running, and the output will be directed to wherever the workflow has specified.

The IDE

A very useful feature of the platform is the Integrated Development Environment (IDE). It is a
command line interface, with a file browser and file editor.

Access the IDE by selecting the icon at the top right of the page.
The following page will be displayed:

At the left is the file browser, at the top right is a file editor, and at the bottom is the command
line. From here, you can issue any linux commands without the need for a separate command
line window.

Note that the file browser is set to the pw folder, which the platform has created in your
home directory. To access any directory outside of this, go to the top right of the file browser
and select the three dots icon. This will bring up a dialog box to select other folders to add to
the workspace. You are free to add files and directories as needed in the pw folder;
everything here is within your AECM home directory.

All workflow code is stored in the directories inside /pw:

 For each workflow you have in the Workflows tab, you will have a workflow directory
in /pw/workflows/;
 The logs for each workflow run are stored in pw/jobs.
 The output for a given workflow run is generally sent to the job directory, but this
default can often be overridden in the workflow parameters.

Further Information

You are now able to run workflows via the web portal. If you have any problems doing so,
please contact support at https://itsupport.einsteinmed.edu/

There are also short videos and other guides available at the Einstein HPC website.

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy