hands on: ai/ml plus hpc€¦ · devops and sysadmins – know their systems and infrastructure –...
Post on 30-Dec-2020
6 Views
Preview:
TRANSCRIPT
1
HANDS ON: AI/ML plus HPC
Craig Gardner, Technical Training ArchitectMember of OpenHPC Technical Steering CommitteeCraig.Gardner@suse.com
Alessandro Festa, Senior Product ManagerTechnical Specialist, AI+MLAlessandro.Festa@suse.com
2
Perfect Marriage
Nothing works as well together as:
Machine Learning
Artificial General Intelligence
High Performance Enterprise Computing
3
Demonstration Objectives
● We can’t make you data scientists
● We can expose you to industry standards and
movement
● We will give you some interesting experience
● You will have a glimpse into what SUSE can do
for you
4
Demonstration Agenda
1) Short Introduction
2) Setup SUSE HPC environment (in an AWS Cloud)
3) Explore MPICH for distributed workloads in HPC
4) Setup JupyterHub in the HPC Cloud
5) Scale the Jupyter Notebook workload with SUSE
HPC
5
Short Introduction
6
Conflict in the Enterprise
DevOps and SysAdmins
– Know their systems and infrastructure
– Know nothing about Data Science
Data Scientists
Know all about their data, models, and tools
– Know just enough about the systems to be dangerous
7
Narrow the Gap
Reduce the headaches for DevOps
Provide an easy environment for Data Scientists
This Demonstration:
– HPC and AI/ML are far bigger than can be fully demonstrated
– This demo is just a glimpse into the Perfect Marriage
8
Setup HPC
9
Setup SUSE HPC environment (in an AWS Cloud)
● Show AWS servers
– 1 head + 4 compute, all running SLES 15 SP1
● Install HPC Module for SLES
● Setup NIS/NFS
● Create a data scientist user
10
Explore MPICH
11
Explore MPICH for distributed workloads in HPC
● Message Passing Interface for HPC
● There are other options for workloads
– openmpi, slurm, mvapich
● In this demo, simply give compute nodes some
work via MPI
12
JupyterHub
13
Setup JupyterHub in the HPC Cloud
● JupyterHub: broad collaboration tool for
Jupyter Notebooks
– Jupyter Notebooks: notations for repeating data science through live
code, equations, and visualizations, guided by inline instructions
● Install and Configure JupyterHub
– Runs on the HPC head node
– Largely based on Python, but other tools expand its functionality, too
– iPython facilitates the means of distributing work
14
Jupyter Notebook in HPC
15
Scale the Jupyter Notebook workload with SUSE HPC
● Setup standard user, called scientist
● As scientist, run Jupyter Notebook
● As scientist, run Jupyter Notebook with
distributed MPICH and HPC resources
16
17
Unpublished Work of SUSE LLC. All Rights Reserved.
This work is an unpublished work and contains confidential, proprietary and trade secret information of SUSE LLC. Access to this work is restricted to SUSE employees who have a need to know to perform tasks within the scope of their assignments. No part of this work may be practiced, performed, copied, distributed, revised, modified, translated, abridged, condensed, expanded, collected, or adapted without the prior written consent of SUSE. Any use or exploitation of this work without authorization could subject the perpetrator to criminal and civil liability.
General Disclaimer
This document is not to be construed as a promise by any participating company to develop, deliver, or market a product. It is not a commitment to deliver any material, code, or functionality, and should not be relied upon in making purchasing decisions. SUSE makes no representations or warranties with respect to the contents of this document, and specifically disclaims any express or implied warranties of merchantability or fitness for any particular purpose. The development, release, and timing of features or functionality described for SUSE products remains at the sole discretion of SUSE. Further, SUSE reserves the right to revise this document and to make changes to its content, at any time, without obligation to notify any person or entity of such revisions or changes. All SUSE marks referenced in this presentation are trademarks or registered trademarks of Novell, Inc. in the United States and other countries. All third-party trademarks are the property of their respective owners.
18
top related