Lecture4 - Performance Analysis CSE 501 Lecture 4 April 8,...

Info iconThis preview shows pages 1–7. Sign up to view the full content.

View Full Document Right Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Performance Analysis CSE 501 Lecture 4 April 8, 2009 Questions? About the last lecture? About the programming assignments? About the reading? Any critiques ready? Why performance analysis? When we write optimizers, we expect to change the performance of our programs. Faster or slower? Bigger or smaller? Without measurement, we're wasting our time. Convincing Performance analysis aims to be convincing . Who are we trying to convince? Ourselves Our bosses Program committees Colleagues We should be so simultaneously curious and skeptical that nobody else will have any unanswered questions. An Example This past year, I built a new code generator for gcc (I had help), based on some new ideas. Eventually, we start to wonder: Is the new code generator any better than the old one? I spent a quarter working on the performance analysis. Luckily, the compiler group at Google already had a good collection of benchmarks, or it would have taken longer. Thorough performance analysis takes time. My Experiments I compiled each program with the old code generator and...
View Full Document

Page1 / 17

Lecture4 - Performance Analysis CSE 501 Lecture 4 April 8,...

This preview shows document pages 1 - 7. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online