1 / 4

Gedae Runtime Kernel Performance Characterization

HPEC 2003. Gedae Runtime Kernel Performance Characterization. Kerry Barnes, William Lundgren, James Steed Gedae, Inc. Efficiency of Gedae Applications. How efficient are Gedae applications compared to hand coding? What is the overhead of the Gedae Runtime Kernel?

keegan
Download Presentation

Gedae Runtime Kernel Performance Characterization

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. HPEC 2003 Gedae Runtime Kernel Performance Characterization Kerry Barnes, William Lundgren, James Steed Gedae, Inc. Gedae, Inc. www.gedae.com

  2. Efficiency of Gedae Applications • How efficient are Gedae applications compared to hand coding? • What is the overhead of the Gedae Runtime Kernel? • What is the cost of executing a primitive? • What is the cost of dynamic vs. static scheduling? Gedae, Inc. www.gedae.com

  3. Instrumenting Kernel • Gedae Kernel instrumented with timer functions • 44 sections of code were timed • Benchmarks • rt_stay_easy: static scheduling • e_comm: dynamic scheduling • noise_removal: dynamic scheduling with segmentation • Timings gathered on AltiVec targets from 3 vendors • 1-3 usec overhead for all benchmarks Gedae, Inc. www.gedae.com

  4. Function rt_stap_ easy e_comm noise_ removal run-sched-pass 90% 20% 18% copy functions 18% 9% schedule state 36% 26% queue manage-ment 10% 15% Total % of Overhead 90% 84% 68% Categorize Overhead • Four categories accounted for most of the overhead time • Suggested enhancements: • Code generation of static schedule • Eliminate copies by directly adjusting pointers • Improve schedule state change functions We can comprehensively measure the overhead and systematically optimize the most costly components Gedae, Inc. www.gedae.com

More Related