This preview shows pages 1–2. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.View Full Document
Unformatted text preview: Programming for Performance: Assignment 3 (v1) Patrick Lam Due: March 7, 2011 Now that youve manually written parallel code using pthreads or a glib interface to it, well move on to 1) automatic parallelization (L12) and 2) manu- ally specifying parallelization with OpenMP (L13 v2), both 2a) for array-based programs (for-loop parallelization) and 2b) task-based OpenMP parallelization. Gentle reminder: you need to hand in code that compiles. Otherwise you get 0 points for that question. Part 1: automatic parallelization (40 points) In this part of the assignment, youll play with automatic parallelization and try to get it working on some production code. In class, I talked about the Oracle Solaris Studio compiler, which you can download on the Internet. I also found another free (as-in-speech) parallelizing compiler, Open64, which AMD productized 1 . You can read the documentation 2 and find out about the-apo (automatic parallelization) and-mso (optimize for multicores) options. Figure(optimize for multicores) options....
View Full Document
- Spring '11