Introduction to Supercomputing Systems

You are not logged in.
Please Log In for full access to the web site.
Note that this link will take you to an external site (https://petrock.mit.edu) to authenticate, and then you will be redirected back to this page.

Check your knowledge of the concepts from the Introduction to Supercomputing Systems module

1) Introduction and Systems Review

True or False: HPC Systems are made up of the same or similar components as a laptop or desktop, but on a larger scale.

True

False

Which component is responsible for computation in a computer or HPC System?

Processor

Storage

Network

Operating Systems/Scheduler

RAM/Memory is an example of

Processor

Storage

Network

Operating Systems/Scheduler

Which component enables internet access

Processor

Storage

Network

Operating Systems/Scheduler

Which is responsible for determining where an application runs (on what hardware)

Processor

Storage

Network

Operating Systems/Scheduler

In an HPC System storage is:

Local to each node and only accessible by the node

Shared and can be accessed by all nodes

Both local and shared. Each Node has its own local storage and can also access shared storage

True or False: RAM (Memory) on one compute node in an HPC System is accessible from other compute nodes.

True

False

Processes on different compute nodes can communicate over the:

RAM

Network

CPUs or GPUs

True or False: Users decide which specific compute nodes their applications run on in an HPC System.

True

False

2) Scheduler Knowledge Check

Appropriate use of the login node includes: (Check all that apply)

Editing code/text files

Running code

Staging data

Running small test cases

Downloading data and packages

The Scheduler represents a High Performance Computing System component not found in desktop systems. The use of a scheduler insures that (Check all that apply)

Jobs submitted to the system are using the number of processors required for the job

Jobs submitted to the system have access to the memory needed for the job

Users don't need to monitor the number of available processors in order to launch their jobs

Multiple concurrent user jobs can be accommodated in a fair way (no user monopolizes the resources)

True or False: If I am not running in parallel, I do not need to use the scheduler.

True

False