Supercomputing Systems Deep Dive

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 Supercomputing Systems Deep Dive module

1) Processor

What is a processor

The processor interacts with what to fetch instructions and store results

Choose the tasks that can be done on a CPU

Choose the tasks that can be done on a CPU

2) Networks

InfiniBand, OmniPath, and ethernet are examples of:

Which networking device can send messages across defined network boundaries?

Which networking device broadcasts the same message to all machines that are connected to it?

Which networking device can send a message to the machine on a specific port, but cannot send messages across defined network boundaries?

Which networking topologies are most commonly used in HPC Systems? (Check all that apply)

Why is WiFi or other wireless networks not used for HPC Systems?

Which stores data on rotating magnetic platters?

Which stores data in transistor circuits on chips

What is responsible for managing the location of data objects in the Lustre high performance filesystem?

True or False: The HPC application requests a file directly from the Lustre metatdata server and fetches it from the data servers.

3) Resource Management

True or False: HPC Systems have a scheduler (or resource manager), and so do not need an operating system.

The Scheduler is responsible for:

True or False: HPC Systems have a scheduler (or resource manager), and so do not need an operating system.

Job execution (dispatching, monitoring, and retiring running jobs) is handled by: