Kostas Papadopoulos, Kyriakos Stavrou, P. Trancoso
{"title":"HelperCore_DB: Exploiting Multicore Technology for Databases","authors":"Kostas Papadopoulos, Kyriakos Stavrou, P. Trancoso","doi":"10.1109/PACT.2007.37","DOIUrl":"https://doi.org/10.1109/PACT.2007.37","url":null,"abstract":"The current trend in microprocessor design is to keep the design simple while packing more processors on the same die in order to increase the performance. These chips are called chip multiprocessors (CMP) or multicores. While currently the parallelism provided by the multiple cores is used to achieve higher throughput, it is possible to split a single application into multiple tasks and execute them in parallel. In this work we propose to exploit the cores of a multi-core architecture in an alternative way. Instead of executing application code on all cores, some cores execute code that will indirectly result in a performance benefit for the application. We call these the helper cores and for this work we focus on the implementation of HelperCoreDB, a helper core for improving the performance of database workloads. The proposed implementation focuses on solving one of the most serious problems in memory-resident database applications, the memory latency penalty, by performing efficient data prefetching. As the prefetching is done by the helper core and the data is used in a different core, the only hardware requirement is that the helper and application cores share some level of the memory hierarchy.","PeriodicalId":401396,"journal":{"name":"16th International Conference on Parallel Architecture and Compilation Techniques (PACT 2007)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115658541","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}
{"title":"Fast Track: Supporting Unsafe Optimizations with Software Speculation","authors":"Kirk Kelsey, Chengliang Zhang, C. Ding","doi":"10.1109/PACT.2007.35","DOIUrl":"https://doi.org/10.1109/PACT.2007.35","url":null,"abstract":"The shift in processor technology toward multi-core, multi-processors opens new opportunities for software speculation where program code is speculatively executed to improve speed at the cost of having handle errors. In this paper we describe a new use of software speculation in a system we call fast track in which we can use some unsafely optimized code while still ensuring correctness.","PeriodicalId":401396,"journal":{"name":"16th International Conference on Parallel Architecture and Compilation Techniques (PACT 2007)","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115661779","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}
{"title":"Bridging Inputs and Program Dynamic Behavior","authors":"Xipeng Shen, Feng Mao","doi":"10.1109/PACT.2007.18","DOIUrl":"https://doi.org/10.1109/PACT.2007.18","url":null,"abstract":"Program optimizations have evolved from static to dynamic. However, runtime optimization often suffers from not knowing global behavior of a program's execution, and not affording sophisticated program analysis. On the other hand, offline profiling techniques overcome both obstacles but are oblivious to the effects of program inputs. This work attempts to address the problem by developing a framework including three components: an extensible Input Characterization Language (XICL), statistical feature selection, and regression model building. The first two components extract important attributes from program inputs, and the last one builds models between those attributes and program dynamic behavior.","PeriodicalId":401396,"journal":{"name":"16th International Conference on Parallel Architecture and Compilation Techniques (PACT 2007)","volume":"12 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127455143","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}
{"title":"A Phase-Adaptive Approach to Increasing Cache Performance","authors":"Matthew A. Watkins, S. Mckee, Lambert Schaelicke","doi":"10.1109/PACT.2007.9","DOIUrl":"https://doi.org/10.1109/PACT.2007.9","url":null,"abstract":"Technological advances along with more complex and dynamic application behavior argue for revisiting mechanisms that adapt logical cache block size to application characteristics. This approach to bridging the processor/memory performance gap has been studied in the past, but most studies used trace-driven simulation and only looked at L1 caches. Given the changes in hardware and software since these seminal studies, we revisit the general approach: we present a transparent, phase-adaptive mechanism for L2 cache block superloading with minimal hardware complexity, evaluating it on a full-system simulator running 23 SPEC CPU2000 applications run to completion using training inputs.","PeriodicalId":401396,"journal":{"name":"16th International Conference on Parallel Architecture and Compilation Techniques (PACT 2007)","volume":"15 4","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114039228","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}
Javier Vera, F. Cazorla, Alex Pajuelo, Oliverio J. Santana, E. Fernández, M. Valero
{"title":"FAME: FAirly MEasuring Multithreaded Architectures","authors":"Javier Vera, F. Cazorla, Alex Pajuelo, Oliverio J. Santana, E. Fernández, M. Valero","doi":"10.1109/PACT.2007.33","DOIUrl":"https://doi.org/10.1109/PACT.2007.33","url":null,"abstract":"Nowadays, multithreaded architectures are becoming more and more popular. In order to evaluate their behavior, several methodologies and metrics have been proposed. A methodology defines when the measurements of a given workload execution are taken. A metric combines those measurements to obtain a final evaluation result. However, since current evaluation methodologies do not provide representative measurements for these metrics, the analysis and evaluation of novel ideas could be either unfair or misleading. Given the potential impact of multithreaded architectures on current and future processor designs, it is crucial to develop an accurate evaluation methodology for them. This paper presents FAME, a new evaluation methodology aimed to fairly measure the performance of multithreaded processors. FAME re-executes all traces in a multithreaded workload until all of them are fairly represented in the final measurements taken from the workload. We compare FAME with previously used methodologies for both architectural research simulators and real processors. Our results show that FAME provides more accurate measurements than other methodologies, becoming an ideal evaluation methodology to analyze proposals for multithreaded architectures.","PeriodicalId":401396,"journal":{"name":"16th International Conference on Parallel Architecture and Compilation Techniques (PACT 2007)","volume":"115 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130032410","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}
Georgiana Macariu, M. Frîncu, A. Cârstea, D. Petcu, A. Eckstein
{"title":"Redesigning Parallel Symbolic Computations Packages","authors":"Georgiana Macariu, M. Frîncu, A. Cârstea, D. Petcu, A. Eckstein","doi":"10.1109/PACT.2007.57","DOIUrl":"https://doi.org/10.1109/PACT.2007.57","url":null,"abstract":"The ability of multi-core processors to increase application performance depends on the use of multiple threads within applications. Symbolic computations, requiring both CPU power and large memory, are well-suited candidates for deriving advantages from multi-core parallel architectures. This is possible only if the specific libraries and tools are designed to allow multi-threading and multi-processes. In order to promote the changes needed to adapt these libraries and tools to the new architectures, the changes performed on two main algorithms for symbolic computations, parallel integer factorization and parallel Grobner base computation, are described and discussed.","PeriodicalId":401396,"journal":{"name":"16th International Conference on Parallel Architecture and Compilation Techniques (PACT 2007)","volume":"21 2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132266225","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}
{"title":"Unified Architectural Support for Soft-Error Protection or Software Bug Detection","authors":"Martin Dimitrov, Huiyang Zhou","doi":"10.1109/PACT.2007.76","DOIUrl":"https://doi.org/10.1109/PACT.2007.76","url":null,"abstract":"In this paper we propose a unified architectural support that can be used flexibly for either soft-error protection or software bug detection. Our approach is based on dynamically detecting and enforcing instruction- level invariants. A hardware table is designed to keep track of run-time invariant information. During program execution, instructions access this table and compare their produced results against the stored invariants. Any violation of the predicted invariant suggests a potential abnormal behavior, which could be a result of a soft error or a latent software bug. In case of a soft error, monitoring invariant violations provides opportunistic soft-error protection to multiple structures in processor pipelines. Our experimental results show that invariant violations detect soft errors promptly and as a result, simple pipeline squashing is able to fix most of the detected soft errors. Meanwhile, the same approach can be easily adapted for software bug detection. The proposed architectural support eliminates the substantial performance overhead associated with software-based bug-detection approaches and enables continuous monitoring of production code.","PeriodicalId":401396,"journal":{"name":"16th International Conference on Parallel Architecture and Compilation Techniques (PACT 2007)","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133123393","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}
J. H. Kelm, Isaac Gelado, M. Murphy, N. Navarro, S. Lumetta, W. Hwu
{"title":"CIGAR: Application Partitioning for a CPU/Coprocessor Architecture","authors":"J. H. Kelm, Isaac Gelado, M. Murphy, N. Navarro, S. Lumetta, W. Hwu","doi":"10.1109/PACT.2007.21","DOIUrl":"https://doi.org/10.1109/PACT.2007.21","url":null,"abstract":"We present CIGAR, a methodology and development platform that facilitates the use of data-parallel coprocessors. With CIGAR, application developers use profiling tools to identify parts of the application for data-parallel execution, determine the application data structures to be hosted by the coprocessor, prototype coprocessor execution of these parts, and debug correctness of partitioned execution of the application using emulation. The CIGAR methodology is complemented by a CPU/FPGA prototyping platform that runs a fully functional version of the Linux operating system and associated development tools and libraries. To guide the development of our work and to evaluate its utility, we have instrumented SPECint2006 applications to utilize coprocessors emulated by softcore processors embedded in our prototyping platform. Examples of how a developer would use CIGAR to partition an application for a heterogeneous CPU/coprocessor environment are demonstrated.","PeriodicalId":401396,"journal":{"name":"16th International Conference on Parallel Architecture and Compilation Techniques (PACT 2007)","volume":"46 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114803282","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}
{"title":"Dynamic Cache Placement with Two-level Mapping to Reduce Conflict Misses","authors":"K. Rajan, R. Govindarajan, B. Amrutur","doi":"10.1109/PACT.2007.27","DOIUrl":"https://doi.org/10.1109/PACT.2007.27","url":null,"abstract":"Due to the tight coupling between processor cycle time and L1 access time, L1 caches are typically small and have low associativities. As a consequence they incur a higher percentage of conflict misses than lower level caches. The extent of conflict depends on the memory access pattern exhibited by the program, and can vary from program to program. By using a fixed set of bits to index the cache, conventional mapping enforces the same rigid mapping from address to cache set for all programs. This results in a non-uniform distribution of addresses among cache sets causing unnecessary conflict misses. Such conflicts could be avoided if some flexibility in mapping is exercised.","PeriodicalId":401396,"journal":{"name":"16th International Conference on Parallel Architecture and Compilation Techniques (PACT 2007)","volume":"38 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129758704","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}
{"title":"Hybrid Specialization: A Trade-off Between Static and Dynamic Specialization","authors":"M. Khan, H. Charles, Denis Barthou","doi":"10.1109/PACT.2007.38","DOIUrl":"https://doi.org/10.1109/PACT.2007.38","url":null,"abstract":"Code specialization is a well-known technique used to produce more efficient code from a generic one. This technique is widely used, from high level languages to optimized libraries like FFTW. It works by substituting a formal input value by an effective value, and can be done either statically or dynamically. Static specialization makes use of data that is expected to be frequently used, whereas, dynamic specialization uses the actual values at run-time. We propose a novel hybrid method combining both the static and the dynamic specialization.","PeriodicalId":401396,"journal":{"name":"16th International Conference on Parallel Architecture and Compilation Techniques (PACT 2007)","volume":"34 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132438352","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}