|
What is High Performance Computing (HPC)
|
High Performance Computing (HPC) typically involves connecting to very large computing systems elsewhere in the world.
These systems can be used to do work that would either be impossible or much slower on smaller systems.
|
|
Connecting to a remote HPC System
|
|
|
Deep learning CPU vs GPU
|
Open On Demand requires you have a strong, stable internet connection whereas SSH can work with weak connections too.
JupyterLab is a more common platform for data science research but there are other IDE (Integrated Development Environment softwares) like PyCharm, Spyder, RMarkdown too.
Using multiple GPUs won’t improve the performance of your machine learning model. It only helps for a very complex computation or large models.
|
|
Staging and File System Choice
|
File system is a way in which operating systems organize data on storage devices.
Distributed file system organizes data accross network attached storage devices.
Distributed file system has the advantage of supporting larger scale storage capacity.
Mana supports lustre and NFS file systems.
Lustre on Mana is setup with solid state drives.
NFS on Mana is setup with spinning drives.
|