In this lab, you will be optimizing a program to achieve maximum performance.
This lab is described by a writeup that is available in either PS or PDF format. The associated support files are available in a tar file that you must download and unpack. Further instructions are in the writeup. Since different x86 processors ahve varying performance, you should do the lab on Wilkes, since that's where the grading will be done. (If you optimize on a different box, you might actually get worse results on Wilkes.)
Note: if you are starting this lab before learning
about cache optimization, you should begin by working on the
smooth function. In fact, the
function is probably easier to optimize anyway, despite being
described second in the handout.
© 2005, Geoff Kuenning
This page is maintained by Geoff Kuenning.