DMCA
Harmony: Collection and Analysis of Parallel Block Vectors
| Citations: | 5 - 2 self |
BibTeX
@MISC{Kambadur_harmony:collection,
author = {Melanie Kambadur and Kui Tang and Martha A. Kim},
title = {Harmony: Collection and Analysis of Parallel Block Vectors},
year = {}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
Efficient execution of well-parallelized applications is central to performance in the multicore era. Program analysis tools support the hardware and software sides of this effort by exposing relevant features of multithreaded applications. This paper describes parallel block vectors, which uncover previously unseen characteristics of parallel programs. Parallel block vectors provide block execution profiles per concurrency phase (e.g., the block execution profile of all serial regions of a program). This information provides a direct and fine-grained mapping between an application’s runtime parallel phases and the static code that makes up those phases. This paper also demonstrates how to collect parallel block vectors with minimal application perturbation using Harmony. Harmony is an instrumentation pass for the LLVM compiler that introduces just 16-21 % overhead on average across eight Parsec benchmarks. We apply parallel block vectors to uncover several novel insights about parallel applications with direct consequences for architectural design. First, that the serial and parallel phases of execution used in Amdahl’s Law are often composed of many of the same basic blocks. Second, that program features, such as instruction mix, vary based on the degree of parallelism, with serial phases in particular displaying different instruction mixes from the program as a whole. Third, that dynamic execution frequencies do not necessarily correlate with a block’s parallelism. 1
Keyphrases
parallel block vector block execution profile unseen characteristic fine-grained mapping program analysis tool efficient execution direct consequence well-parallelized application serial region software side concurrency phase multicore era parsec benchmark llvm compiler instruction mix several novel insight instrumentation pas basic block application runtime parallel phase parallel program parallel phase dynamic execution frequency relevant feature architectural design multithreaded application amdahl law block parallelism 16-21 overhead static code different instruction mix minimal application perturbation program feature parallel application serial phase
