2012 21st International Conference on Parallel Architectures and Compilation Techniques (PACT)最新文献

筛选

英文中文

Chrysalis analysis: Incorporating synchronization arcs in dataflow-analysis-based parallel monitoring 蛹分析:在基于数据流分析的并行监控中结合同步弧

2012 21st International Conference on Parallel Architectures and Compilation Techniques (PACT) Pub Date : 2012-09-19 DOI: 10.1145/2370816.2370847

Michelle L. Goodstein, Shimin Chen, Phillip B. Gibbons, M. Kozuch, T. Mowry

{"title":"Chrysalis analysis: Incorporating synchronization arcs in dataflow-analysis-based parallel monitoring","authors":"Michelle L. Goodstein, Shimin Chen, Phillip B. Gibbons, M. Kozuch, T. Mowry","doi":"10.1145/2370816.2370847","DOIUrl":"https://doi.org/10.1145/2370816.2370847","url":null,"abstract":"Software lifeguards, or tools that monitor applications at runtime, are an effective way of identifying program errors and security exploits. Parallel programs are susceptible to a wider range of possible errors than sequential programs, making them even more in need of online monitoring. Unfortunately, monitoring parallel applications is difficult due to inter-thread data dependences. In prior work, we introduced a new software framework for online parallel program monitoring inspired by dataflow analysis, called Butterfly Analysis. Butterfly Analysis uses bounded windows of uncertainty to model the finite upper bound on delay between when an instruction is issued and when all its effects are visible throughout the system. While Butterfly Analysis offers many advantages, it ignored one key source of ordering information which affected its false positive rate: explicit software synchronization, and the corresponding high-level happens-before arcs. In this work we introduce Chrysalis Analysis, which extends the Butterfly Analysis framework to incorporate explicit happens-before arcs resulting from high-level synchronization within a monitored program. We show how to adapt two standard dataflow analysis techniques and two memory and security lifeguards to Chrysalis Analysis, using novel techniques for dealing with the many complexities introduced by happens-before arcs. Our security tool implementation shows that Chrysalis Analysis matches the key advantages of Butterfly Analysis—parallel monitoring, no detailed inter-thread data dependence tracking, no strong memory consistency requirements, and no missed errors—while significantly reducing the number of false positives.","PeriodicalId":403855,"journal":{"name":"2012 21st International Conference on Parallel Architectures and Compilation Techniques (PACT)","volume":"55 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2012-09-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115235822","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 7

Lossless and lossy memory I/O link compression for improving performance of GPGPU workloads 无损和有损内存I/O链路压缩，提高GPGPU工作负载的性能

2012 21st International Conference on Parallel Architectures and Compilation Techniques (PACT) Pub Date : 2012-09-19 DOI: 10.1145/2370816.2370864

V. Sathish, M. Schulte, N. Kim

{"title":"Lossless and lossy memory I/O link compression for improving performance of GPGPU workloads","authors":"V. Sathish, M. Schulte, N. Kim","doi":"10.1145/2370816.2370864","DOIUrl":"https://doi.org/10.1145/2370816.2370864","url":null,"abstract":"State-of-the-art graphic processing units (GPUs) provide very high memory bandwidth, but the performance of many general-purpose GPU (GPGPU) workloads is still bounded by memory bandwidth. Although compression techniques have been adopted by commercial GPUs, they are only used for compressing texture and color data, not data for GPGPU workloads. Furthermore, the microarchitectural details of GPU compression are proprietary and its performance benefits have not been previously published. In this paper, we first investigate required microarchitectural changes to support lossless compression techniques for data transferred between the GPU and its off-chip memory to provide higher effective bandwidth. Second, by exploiting some characteristics of floating-point numbers in many GPGPU workloads, we propose to apply lossless compression to floating-point numbers after truncating their least-significant bits (i.e., lossy compression). This can reduce the bandwidth usage even further with very little impact on overall computational accuracy. Finally, we demonstrate that a GPU with our lossless and lossy compression techniques can improve the performance of memory-bound GPGPU workloads by 26% and 41% on average.","PeriodicalId":403855,"journal":{"name":"2012 21st International Conference on Parallel Architectures and Compilation Techniques (PACT)","volume":"34 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2012-09-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122237021","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 87

Speculative parallelization needs rigor 推测并行化需要严格

2012 21st International Conference on Parallel Architectures and Compilation Techniques (PACT) Pub Date : 2012-09-01 DOI: 10.1145/2370816.2370882

Zhijia Zhao, Bo Wu, Xipeng Shen

引用次数: 0

首页上一页