What is a teraflop?

A teraflop is one trillion floating point operations per second. This measure of compute capacity describes how many multiplications can be performed within one second. Computer vendors and computer resource providers typically list the computing performance of their systems in terms of the number of gigaflops (billion floating point operations per second) or teraflops achieved on standard benchmark programs, such as the LINPACK DP, TPP, and HPC benchmarks, and the SPEC integer and floating point benchmarks. The theoretical peak computing capacity is obtained by multiplying the number of processors by the clock speed of the processor by the number of floating point operations per second that processor type is capable of performing.

"TeraFLOPS" is also sometimes used to indicate one trillion floating point operations per second.

This document was developed with support from the National Science Foundation (NSF) under Grant No. 0503697 to the University of Chicago and subcontracted to Indiana University. Additional support was provided by IU through its participation in the TeraGrid, which is supported by the NSF under Grants No. 0833618, SCI451237, SCI535258, and SCI504075. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the NSF.

This document was developed with support from National Science Foundation (NSF) grant OCI-1053575. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the NSF.

Comments/Questions/Corrections