General Software Information

Here you can find information on using Ohio Supercomputer Center (OSC) software and on our charging policies.

Request for Software Form - This form is used to request that new software be acquired for use at the Ohio Supercomputer Center. (This document is available as an Adobe PDF file.)

Forms

Academic Agreement forms are required to use some third-party software. These documents apply to more than one platform or to no particular platform. A complete list is available on the forms page.

Charging

The following text describes the charging algorithms for academic use on OSC hardware.

Charging for Academic Use

Supercomputer resources on The Cluster are charged in terms of Resource Units, a common measure of the use of CPU and file storage across all platforms.

The charging formula for RUs consumed by a process is:
RU = 0.1 × Tt
where RU is the charge in Resource Units for the process being considered and Tt is the total CPU time in hours for all CPUs.  

For Oakley, the formula for calculating total CPU time differs between parallel and serial jobs.  Parallel jobs are charged for whole nodes regardless of the ppn request submitted.  This means if you were to only request ppn=2 for a 2 node parallel job on Oakley, you would be charged as if you requested a ppn=12 regardless.

For serial jobs on Oakley:

 total CPU time = [walltime used] x [ppn requested]

For Ruby and parallel jobs on Oakley:

 total CPU time = [walltime used] x [max ppn on system] x [nodes requested]

 

Charging on Owens vs Oakley and Ruby

With the introduction of our newest cluster, Owens, we have adjusted our rates to provide a discount to users on Ruby and Owens. The rate for Owens will be the standard of:
RU = 0.1 × Tt

While the rate on Ruby and Oakley will be discounted to:

RU = 0.05 × Tt

This comes out to be a 50% discount with 1 RU allowing for 20 CPU hours on Ruby and Oakley while 1 RU on Owens will be the standard of 10 CPU hrs.

 

Charging for Memory Use

Each processor has a default amount of memory paired along with it. On Oakley this is 4gb/processor. When requesting a specifc amount of memory that doesn't correlate with the default pairing, the charging is as follows:

A job that requests nodes=1:ppn=3 will still be charged for 3 cores worth of Resource Units (RU). However, a job that requests nodes=1:ppn=1,mem=12GB is currently only charged for 1 core worth of RU, in accordance with our current charging policy. Under our new policy, this job would be charged for 3 cores. Why?

effective cores = memory / memory per core

or

effective cores = 12GB / (4GB/core) = 3 core

The value for "effective cores" will be used in the regular RU calculation if larger than the explicit number of cores requested.

Jobs requesting more than 48GB memory are allocated (and charged for) an entire large-memory node.

Please note that this change has no charging impact on jobs that use full nodes, including parallel jobs.

 

GPU Usage

There is no charge specifically for GPU usage.   There is a difference in how GPU jobs are allocated normal processors, and therefore charged, between Oakley and Glenn.  Serial GPU jobs on Oakley may request partial nodes ( ppn < 12 ) and will only be allocated their requested processors like a normal serial job.

 

File Storage
Charging for Academic Use

NOTE: At this time, there is no charge for file storage.

Statewide Software Library

The Ohio Supercomputer Center manages a program that offers state-of-the art software installed at the user's home site. Originally this program was titled "Statewide Software Library." That is subsumed by the later program: Statewide Software Distribution.