Chapter 25

Trinity workloads

Abstract

Explores Trinity mini-applications that were used in two early Knights Landing procurements. Discusses the performance impact of different memory modes, cluster modes, and problem sizes on overall performance. Results for all workloads are presented; a detailed analysis of miniGhost, one of the workloads, is presented including the software optimizations to improve its performance.

Keywords

Trinity; MCDRAM; MiniGhost; MiniFE; SNAP; GTC; UMT; AMG; MiniDFT; Problem sizing; Weak scaling; Cache unfriendly; Cache friendly; Broadwell; MPI

What is new with Knights Landing in this chapter?

Use of MCDRAM, Hyperthreading and AVX-512.

In this chapter, we explore a set of High Performance Computing workloads, the ...

Get Intel Xeon Phi Processor High Performance Programming, 2nd Edition now with the O’Reilly learning platform.

O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.

Start your free trial