Presentation is loading. Please wait.

Presentation is loading. Please wait.

May 10-02 Mike Drob Grant Furgiuele Ben Winters Advisor: Dr. Chris Chu Client: IBM IMB Contact – Karl Erickson.

Similar presentations


Presentation on theme: "May 10-02 Mike Drob Grant Furgiuele Ben Winters Advisor: Dr. Chris Chu Client: IBM IMB Contact – Karl Erickson."— Presentation transcript:

1 May 10-02 Mike Drob Grant Furgiuele Ben Winters Advisor: Dr. Chris Chu Client: IBM IMB Contact – Karl Erickson

2 Project Plan Start with existing serial FastPlace algorithm Parallelize FastPlace algorithm to decrease run-time Hope to gain increases as close to N times speedup (N = cores) as possible Realistically, 0.75N End-goal is mostly proof-of-concept IBM uses in-house algorithm Contains proprietary circuit processing

3 Project Design Written in C Run under Linux using POSIX thread library Consider scalability – 2, 4, 8, etc. cores RQL implementation IBM Concept Netlist optimization for placement Currently implemented, needs more testing

4 Implementation Using Data Parallelism as scheme Assigning loop iterations to threads Especially with double-nested loops Localizing variable usage Where absolutely necessary, using thread synchronization (mutex, etc..)

5 Implementation – Thread Pool Threads are created once at start Various Benefits: Minimizes overhead from thread creation Increases cache performance Allows core scalability – number of threads running can equal cores available

6 Testing Profiled original FastPlace algorithm gprof gives CPU time per function Profiling parallel FastPlace qprof and cprof FastPlace code outputs actual time elapsed Can be used to compare performance Not 100% consistent

7 Testing Test results for correctness script will verify outputs Compare “wire length” results Test results for speedup Will compare both CPU time and actual time Threadpool is tested and working Currently getting results for parallelized function which was most CPU intensive Also testing RQL implementation

8 Upcoming Work Continue to parallelize most CPU intensive functions Run profiling tests each time one is completed Run profiling tests on computers with more than 2 cores Verify correctness of placement Gather test results in a readable manner to use on poster

9 Questions?


Download ppt "May 10-02 Mike Drob Grant Furgiuele Ben Winters Advisor: Dr. Chris Chu Client: IBM IMB Contact – Karl Erickson."

Similar presentations


Ads by Google