Removing architectural bottlenecks to the scalability of speculative parallelization

Speculative thread-level parallelization is a promising way to speed up codes that compilers fail to parallelize. While several speculative parallelization schemes have been proposed for different machine sizes and types of codes, the results so far show that it is hard to deliver scalable speedups....

Full description

Saved in:
Bibliographic Details
Published in:Proceedings 28th Annual International Symposium on Computer Architecture pp. 204 - 215
Main Authors: Prvulovic, M., Garzaran, M.J., Rauchwerger, L., Torrellas, J.
Format: Conference Proceeding
Language:English
Japanese
Published: IEEE 2001
Subjects:
Buffer overflow
Costs
Database systems
Design optimization
Hardware
Microprocessors
Parallel programming
Scalability
Traffic control
Yarn
ISBN:0769511627, 9780769511627
ISSN:1063-6897
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Speculative thread-level parallelization is a promising way to speed up codes that compilers fail to parallelize. While several speculative parallelization schemes have been proposed for different machine sizes and types of codes, the results so far show that it is hard to deliver scalable speedups. Often, the problem is not true dependence violations, but sub-optimal architectural design. Consequently, we attempt to identify and eliminate major architectural bottlenecks that limit the scalability of speculative parallelization. The solutions that we propose are: low-complexity commit in constant time to eliminate the task commit bottleneck, a memory-based overflow area to eliminate stall due to speculative buffer overflow, and exploiting high-level access patterns to minimize speculation-induced traffic. To show that the resulting system is truly scalable, we perform simulations with up to 128 processors. With our optimizations, the speedups for 128 and 64 processors reach 63 and 48, respectively. The average speedup for 64 processors is 32, nearly four times higher than without our optimizations.
ISBN:0769511627
9780769511627
ISSN:1063-6897
DOI:10.1109/ISCA.2001.937450