Both commercial and scientific workloads benefit from concurrency and exhibit data sharing across threads/processes. The resulting sharing patterns are often fine-grain, with the modified cache lines still residing in the writer's primary cache when accessed. Chip multiprocessors present an opportunity to optimize for fine-grain sharing using direct access to remote processor components through low-latency on-chip interconnects. In this paper, we present Adaptive Replication, Migration, and producer-Consumer Optimization (ARMCO), a coherence protocol that, to the best of our knowledge, is the first to exploit direct access to the L1 caches of remote processors (rather than via coherence mechanisms) in order to support fine-grain sharing. Our goal is to provide support for tightly coupled sharing by recognizing and adapting to common sharing patterns such as migratory, producer-consumer, multiple-reader, and multiple read-write. The protocol places data close to where it is most needed and leverages direct access when following conventional coherence actions proves wasteful. Via targeted optimizations for each of these access patterns, our proposed protocol is able to reduce the average access latency and increase the effective cache capacity at the L1 level with on-chip storage overhead as low as 0.38%. Full-system simulations of 16-processor CMPs show an average (geometric mean) speedup of 1.13 (ranging from 1.04 to 2.26) for 12 commercial, scientific, and mining workloads, with an average of 1.18 if we include 2 microbenchmarks. ARMCO also reduces the on-chip bandwidth requirements and dynamic energy (power) consumption by an average of 33.3% and 31.2% (20.2%) respectively. By evaluating optimizations at both the L1 and the L2 level, we demonstrate that when considering performance, optimization at the L1 level is more effective at supporting fine-grain sharing than that at the L2 level.
{"title":"Improving support for locality and fine-grain sharing in chip multiprocessors","authors":"Hemayet Hossain, S. Dwarkadas, Michael C. Huang","doi":"10.1145/1454115.1454138","DOIUrl":"https://doi.org/10.1145/1454115.1454138","url":null,"abstract":"Both commercial and scientific workloads benefit from concurrency and exhibit data sharing across threads/processes. The resulting sharing patterns are often fine-grain, with the modified cache lines still residing in the writer's primary cache when accessed. Chip multiprocessors present an opportunity to optimize for fine-grain sharing using direct access to remote processor components through low-latency on-chip interconnects. In this paper, we present Adaptive Replication, Migration, and producer-Consumer Optimization (ARMCO), a coherence protocol that, to the best of our knowledge, is the first to exploit direct access to the L1 caches of remote processors (rather than via coherence mechanisms) in order to support fine-grain sharing. Our goal is to provide support for tightly coupled sharing by recognizing and adapting to common sharing patterns such as migratory, producer-consumer, multiple-reader, and multiple read-write. The protocol places data close to where it is most needed and leverages direct access when following conventional coherence actions proves wasteful. Via targeted optimizations for each of these access patterns, our proposed protocol is able to reduce the average access latency and increase the effective cache capacity at the L1 level with on-chip storage overhead as low as 0.38%. Full-system simulations of 16-processor CMPs show an average (geometric mean) speedup of 1.13 (ranging from 1.04 to 2.26) for 12 commercial, scientific, and mining workloads, with an average of 1.18 if we include 2 microbenchmarks. ARMCO also reduces the on-chip bandwidth requirements and dynamic energy (power) consumption by an average of 33.3% and 31.2% (20.2%) respectively. By evaluating optimizations at both the L1 and the L2 level, we demonstrate that when considering performance, optimization at the L1 level is more effective at supporting fine-grain sharing than that at the L2 level.","PeriodicalId":186773,"journal":{"name":"2008 International Conference on Parallel Architectures and Compilation Techniques (PACT)","volume":"48 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-10-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124815720","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}
Marc González, Nikola Vujic, X. Martorell, E. Ayguadé, A. Eichenberger, Tong Chen, Zehra Sura, Tao Zhang, K. O'Brien, Kathryn M. O'Brien
Ease of programming is one of the main impediments for the broad acceptance of multi-core systems with no hardware support for transparent data transfer between local and global memories. Software cache is a robust approach to provide the user with a transparent view of the memory architecture; but this software approach can suffer from poor performance. In this paper, we propose a hierarchical, hybrid software-cache architecture that classifies at compile time memory accesses in two classes, high-locality and irregular. Our approach then steers the memory references toward one of two specific cache structures optimized for their respective access pattern. The specific cache structures are optimized to enable high-level compiler optimizations to aggressively unroll loops, reorder cache references, and/or transform surrounding loops so as to practically eliminate the software cache overhead in the innermost loop. Performance evaluation indicates that improvements due to the optimized software-cache structures combined with the proposed code-optimizations translate into 3.5 to 8.4 speedup factors, compared to a traditional software cache approach. As a result, we demonstrate that the Cell BE processor can be a competitive alternative to a modern server-class multi-core such as the IBM Power5 processor for a set of parallel NAS applications.
{"title":"Hybrid access-specific software cache techniques for the cell BE architecture","authors":"Marc González, Nikola Vujic, X. Martorell, E. Ayguadé, A. Eichenberger, Tong Chen, Zehra Sura, Tao Zhang, K. O'Brien, Kathryn M. O'Brien","doi":"10.1145/1454115.1454156","DOIUrl":"https://doi.org/10.1145/1454115.1454156","url":null,"abstract":"Ease of programming is one of the main impediments for the broad acceptance of multi-core systems with no hardware support for transparent data transfer between local and global memories. Software cache is a robust approach to provide the user with a transparent view of the memory architecture; but this software approach can suffer from poor performance. In this paper, we propose a hierarchical, hybrid software-cache architecture that classifies at compile time memory accesses in two classes, high-locality and irregular. Our approach then steers the memory references toward one of two specific cache structures optimized for their respective access pattern. The specific cache structures are optimized to enable high-level compiler optimizations to aggressively unroll loops, reorder cache references, and/or transform surrounding loops so as to practically eliminate the software cache overhead in the innermost loop. Performance evaluation indicates that improvements due to the optimized software-cache structures combined with the proposed code-optimizations translate into 3.5 to 8.4 speedup factors, compared to a traditional software cache approach. As a result, we demonstrate that the Cell BE processor can be a competitive alternative to a modern server-class multi-core such as the IBM Power5 processor for a set of parallel NAS applications.","PeriodicalId":186773,"journal":{"name":"2008 International Conference on Parallel Architectures and Compilation Techniques (PACT)","volume":"55 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-10-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128267827","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}
Technology scaling in integrated circuits has consistently provided dramatic performance improvements in modern microprocessors. However, increasing device counts and decreasing on-chip voltage levels have made transient errors a first-order design constraint that can no longer be ignored. Several proposals have provided fault detection and tolerance through redundantly executing a program on an additional hardware thread or core. While such techniques can provide high fault coverage, they at best provide equivalent performance to the original execution and at worst incur a slowdown due to error checking, contention for shared resources, and synchronization overheads. This work achieves a similar goal of detecting transient errors by redundantly executing a program on an additional processor core, however it speeds up (rather than slows down) program execution compared to the unprotected baseline case. It makes the observation that a small number of instructions are detrimental to overall performance, and selectively skipping them enables one core to advance far ahead of the other to obtain prefetching and large instruction window benefits. We highlight the modest incremental hardware required to support skewed redundancy and demonstrate a speedup of 6%/54% for a collection of integer/floating point benchmarks while still providing 100% error detection coverage within our sphere of replication. Additionally, we show that a third core can further improve performance while adding error recovery capabilities.
{"title":"Skewed redundancy","authors":"Gordon B. Bell, Mikko H. Lipasti","doi":"10.1145/1454115.1454126","DOIUrl":"https://doi.org/10.1145/1454115.1454126","url":null,"abstract":"Technology scaling in integrated circuits has consistently provided dramatic performance improvements in modern microprocessors. However, increasing device counts and decreasing on-chip voltage levels have made transient errors a first-order design constraint that can no longer be ignored. Several proposals have provided fault detection and tolerance through redundantly executing a program on an additional hardware thread or core. While such techniques can provide high fault coverage, they at best provide equivalent performance to the original execution and at worst incur a slowdown due to error checking, contention for shared resources, and synchronization overheads. This work achieves a similar goal of detecting transient errors by redundantly executing a program on an additional processor core, however it speeds up (rather than slows down) program execution compared to the unprotected baseline case. It makes the observation that a small number of instructions are detrimental to overall performance, and selectively skipping them enables one core to advance far ahead of the other to obtain prefetching and large instruction window benefits. We highlight the modest incremental hardware required to support skewed redundancy and demonstrate a speedup of 6%/54% for a collection of integer/floating point benchmarks while still providing 100% error detection coverage within our sphere of replication. Additionally, we show that a third core can further improve performance while adding error recovery capabilities.","PeriodicalId":186773,"journal":{"name":"2008 International Conference on Parallel Architectures and Compilation Techniques (PACT)","volume":"34 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-10-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127620106","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}
Matthew Curtis-Maury, Ankur Shah, F. Blagojevic, Dimitrios S. Nikolopoulos, B. Supinski, M. Schulz
Power has become a primary concern for HPC systems. Dynamic voltage and frequency scaling (DVFS) and dynamic concurrency throttling (DCT) are two software tools (or knobs) for reducing the dynamic power consumption of HPC systems. To date, few works have considered the synergistic integration of DVFS and DCT in performance-constrained systems, and, to the best of our knowledge, no prior research has developed application-aware simultaneous DVFS and DCT controllers in real systems and parallel programming frameworks. We present a multi-dimensional, online performance predictor, which we deploy to address the problem of simultaneous runtime optimization of DVFS and DCT on multi-core systems. We present results from an implementation of the predictor in a runtime library linked to the Intel OpenMP environment and running on an actual dual-processor quad-core system. We show that our predictor derives near-optimal settings of the power-aware program adaptation knobs that we consider. Our overall framework achieves significant reductions in energy (19% mean) and ED2 (40% mean), through simultaneous power savings (6% mean) and performance improvements (14% mean). We also find that our framework outperforms earlier solutions that adapt only DVFS or DCT, as well as one that sequentially applies DCT then DVFS. Further, our results indicate that prediction-based schemes for runtime adaptation compare favorably and typically improve upon heuristic search-based approaches in both performance and energy savings.
{"title":"Prediction models for multi-dimensional power-performance optimization on many cores","authors":"Matthew Curtis-Maury, Ankur Shah, F. Blagojevic, Dimitrios S. Nikolopoulos, B. Supinski, M. Schulz","doi":"10.1145/1454115.1454151","DOIUrl":"https://doi.org/10.1145/1454115.1454151","url":null,"abstract":"Power has become a primary concern for HPC systems. Dynamic voltage and frequency scaling (DVFS) and dynamic concurrency throttling (DCT) are two software tools (or knobs) for reducing the dynamic power consumption of HPC systems. To date, few works have considered the synergistic integration of DVFS and DCT in performance-constrained systems, and, to the best of our knowledge, no prior research has developed application-aware simultaneous DVFS and DCT controllers in real systems and parallel programming frameworks. We present a multi-dimensional, online performance predictor, which we deploy to address the problem of simultaneous runtime optimization of DVFS and DCT on multi-core systems. We present results from an implementation of the predictor in a runtime library linked to the Intel OpenMP environment and running on an actual dual-processor quad-core system. We show that our predictor derives near-optimal settings of the power-aware program adaptation knobs that we consider. Our overall framework achieves significant reductions in energy (19% mean) and ED2 (40% mean), through simultaneous power savings (6% mean) and performance improvements (14% mean). We also find that our framework outperforms earlier solutions that adapt only DVFS or DCT, as well as one that sequentially applies DCT then DVFS. Further, our results indicate that prediction-based schemes for runtime adaptation compare favorably and typically improve upon heuristic search-based approaches in both performance and energy savings.","PeriodicalId":186773,"journal":{"name":"2008 International Conference on Parallel Architectures and Compilation Techniques (PACT)","volume":"413 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-10-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126692646","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}
The emergence of power as a first-class design constraint has fueled the proposal of a growing number of run-time power optimizations. Many of these optimizations trade-off power saving opportunity for a variable performance loss which depends on application characteristics and program phase. Furthermore, the potential benefits of these optimizations are sometimes non-additive, and it can be difficult to identify which combinations of these optimizations to apply. Trial-and-error approaches have been proposed to adaptively tune a processor. However, in a chip multiprocessor, the cost of individually configuring each core under a wide range of optimizations would be prohibitive under simple trial-and-error approaches. In this work, we introduce an adaptive, multi-optimization power saving strategy for multi-core power management. Specifically, we solve the problem of meeting a global chip-wide power budget through run-time adaptation of highly configurable processor cores. Our approach applies analytic modeling to reduce exploration time and decrease the reliance on trial-and-error methods. We also introduce risk evaluation to balance the benefit of various power saving optimizations versus the potential performance loss. Overall, we find that our approach can significantly reduce processor power consumption compared to alternative optimization strategies.
{"title":"Multi-Optimization power management for chip multiprocessors","authors":"Ke Meng, R. Joseph, R. Dick, L. Shang","doi":"10.1145/1454115.1454141","DOIUrl":"https://doi.org/10.1145/1454115.1454141","url":null,"abstract":"The emergence of power as a first-class design constraint has fueled the proposal of a growing number of run-time power optimizations. Many of these optimizations trade-off power saving opportunity for a variable performance loss which depends on application characteristics and program phase. Furthermore, the potential benefits of these optimizations are sometimes non-additive, and it can be difficult to identify which combinations of these optimizations to apply. Trial-and-error approaches have been proposed to adaptively tune a processor. However, in a chip multiprocessor, the cost of individually configuring each core under a wide range of optimizations would be prohibitive under simple trial-and-error approaches. In this work, we introduce an adaptive, multi-optimization power saving strategy for multi-core power management. Specifically, we solve the problem of meeting a global chip-wide power budget through run-time adaptation of highly configurable processor cores. Our approach applies analytic modeling to reduce exploration time and decrease the reliance on trial-and-error methods. We also introduce risk evaluation to balance the benefit of various power saving optimizations versus the potential performance loss. Overall, we find that our approach can significantly reduce processor power consumption compared to alternative optimization strategies.","PeriodicalId":186773,"journal":{"name":"2008 International Conference on Parallel Architectures and Compilation Techniques (PACT)","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-10-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122229590","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}
Bingsheng He, Wenbin Fang, Qiong Luo, N. Govindaraju, Tuyong Wang
We design and implement Mars, a MapReduce framework, on graphics processors (GPUs). MapReduce is a distributed programming framework originally proposed by Google for the ease of development of web search applications on a large number of commodity CPUs. Compared with CPUs, GPUs have an order of magnitude higher computation power and memory bandwidth, but are harder to program since their architectures are designed as a special-purpose co-processor and their programming interfaces are typically for graphics applications. As the first attempt to harness GPU's power for MapReduce, we developed Mars on an NVIDIA G80 GPU, which contains over one hundred processors, and evaluated it in comparison with Phoenix, the state-of-the-art MapReduce framework on multi-core CPUs. Mars hides the programming complexity of the GPU behind the simple and familiar MapReduce interface. It is up to 16 times faster than its CPU-based counterpart for six common web applications on a quad-core machine.
{"title":"Mars: A MapReduce Framework on graphics processors","authors":"Bingsheng He, Wenbin Fang, Qiong Luo, N. Govindaraju, Tuyong Wang","doi":"10.1145/1454115.1454152","DOIUrl":"https://doi.org/10.1145/1454115.1454152","url":null,"abstract":"We design and implement Mars, a MapReduce framework, on graphics processors (GPUs). MapReduce is a distributed programming framework originally proposed by Google for the ease of development of web search applications on a large number of commodity CPUs. Compared with CPUs, GPUs have an order of magnitude higher computation power and memory bandwidth, but are harder to program since their architectures are designed as a special-purpose co-processor and their programming interfaces are typically for graphics applications. As the first attempt to harness GPU's power for MapReduce, we developed Mars on an NVIDIA G80 GPU, which contains over one hundred processors, and evaluated it in comparison with Phoenix, the state-of-the-art MapReduce framework on multi-core CPUs. Mars hides the programming complexity of the GPU behind the simple and familiar MapReduce interface. It is up to 16 times faster than its CPU-based counterpart for six common web applications on a quad-core machine.","PeriodicalId":186773,"journal":{"name":"2008 International Conference on Parallel Architectures and Compilation Techniques (PACT)","volume":"75 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-10-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127675454","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}
Manman Ren, Ji Young Park, M. Houston, A. Aiken, W. Dally
Achieving good performance on a modern machine with a multi-level memory hierarchy, and in particular on a machine with software-managed memories, requires precise tuning of programs to the machine's particular characteristics. A large program on a multi-level machine can easily expose tens or hundreds of inter-dependent parameters which require tuning, and manually searching the resultant large, non-linear space of program parameters is a tedious process of trial-and-error. In this paper we present a general framework for automatically tuning general applications to machines with software-managed memory hierarchies. We evaluate our framework by measuring the performance of benchmarks that are tuned for a range of machines with different memory hierarchy configurations: a cluster of Intel P4 Xeon processors, a single Cell processor, and a cluster of Sony Playstation3's.
{"title":"A tuning framework for software-managed memory hierarchies","authors":"Manman Ren, Ji Young Park, M. Houston, A. Aiken, W. Dally","doi":"10.1145/1454115.1454155","DOIUrl":"https://doi.org/10.1145/1454115.1454155","url":null,"abstract":"Achieving good performance on a modern machine with a multi-level memory hierarchy, and in particular on a machine with software-managed memories, requires precise tuning of programs to the machine's particular characteristics. A large program on a multi-level machine can easily expose tens or hundreds of inter-dependent parameters which require tuning, and manually searching the resultant large, non-linear space of program parameters is a tedious process of trial-and-error. In this paper we present a general framework for automatically tuning general applications to machines with software-managed memory hierarchies. We evaluate our framework by measuring the performance of benchmarks that are tuned for a range of machines with different memory hierarchy configurations: a cluster of Intel P4 Xeon processors, a single Cell processor, and a cluster of Sony Playstation3's.","PeriodicalId":186773,"journal":{"name":"2008 International Conference on Parallel Architectures and Compilation Techniques (PACT)","volume":"11 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-10-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121211565","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}
H. Hayashizaki, Yutaka Sugawara, M. Inaba, K. Hiraki
Massively parallel machines that integrate a large number of simple processors and small scratch-pad memories (SPMs) into a single chip can achieve a high peak performance per watt of power. In these machines, communication optimizations are important because the communication bandwidth tends to be a bottleneck. Previously proposed communication optimizations using copy candidates, which have been shown to be effective, detect frequently reused array regions by compile-time analysis and copy the regions to scratch-pad memories nearer to the processors. However, they have been proposed for uniprocessor systems or small parallel machines with one or more layers of scratch-pad memories, and the analysis time increases when they are applied to massively parallel machines. In this paper, we propose Multilayer Copy-candidate Analysis for Massively Parallel machines (MCAMP), a communication optimization method for massively parallel machines. MCAMP re-formalizes the framework used in earlier works and improves the scalability of the analysis by assuming the homogeneity of the target systems. We implemented an MCAMP optimizer, which takes an input program that consists of perfectly nested loops containing array references and computation codes, and generates optimized communication. We measured the performance of the output programs of the MCAMP optimizer by executing them on a real massively parallel machine GRAPE-DR using a software tool chain that we also implemented. We showed that MCAMP can achieve optimal data transfer patterns and comparable performance to that of hand-optimized codes with a short analysis time.
将大量简单处理器和小型刮刮板存储器(spm)集成到单个芯片中的大规模并行机器可以实现每瓦功率的峰值性能。在这些机器中，通信优化非常重要，因为通信带宽往往是瓶颈。先前提出的使用副本候选的通信优化已被证明是有效的，它通过编译时分析检测频繁重用的数组区域，并将这些区域复制到靠近处理器的临时存储器中。然而，它们已经被提出用于单处理器系统或具有一层或多层刮擦板存储器的小型并行机器，并且当它们应用于大规模并行机器时，分析时间增加。本文提出了一种大规模并行机通信优化方法MCAMP (Multilayer Copy-candidate Analysis for Massively Parallel machine)。MCAMP重新形式化了早期工作中使用的框架，并通过假设目标系统的同质性来提高分析的可伸缩性。我们实现了一个MCAMP优化器，它接受一个由包含数组引用和计算代码的完美嵌套循环组成的输入程序，并生成优化的通信。我们通过在真正的大规模并行机GRAPE-DR上执行MCAMP优化器输出程序来测量它们的性能，并使用了我们实现的软件工具链。我们证明MCAMP可以在较短的分析时间内实现最佳的数据传输模式和与手动优化代码相当的性能。
{"title":"MCAMP: Communication optimization on Massively Parallel Machines with hierarchical scratch-pad memory","authors":"H. Hayashizaki, Yutaka Sugawara, M. Inaba, K. Hiraki","doi":"10.1145/1454115.1454132","DOIUrl":"https://doi.org/10.1145/1454115.1454132","url":null,"abstract":"Massively parallel machines that integrate a large number of simple processors and small scratch-pad memories (SPMs) into a single chip can achieve a high peak performance per watt of power. In these machines, communication optimizations are important because the communication bandwidth tends to be a bottleneck. Previously proposed communication optimizations using copy candidates, which have been shown to be effective, detect frequently reused array regions by compile-time analysis and copy the regions to scratch-pad memories nearer to the processors. However, they have been proposed for uniprocessor systems or small parallel machines with one or more layers of scratch-pad memories, and the analysis time increases when they are applied to massively parallel machines. In this paper, we propose Multilayer Copy-candidate Analysis for Massively Parallel machines (MCAMP), a communication optimization method for massively parallel machines. MCAMP re-formalizes the framework used in earlier works and improves the scalability of the analysis by assuming the homogeneity of the target systems. We implemented an MCAMP optimizer, which takes an input program that consists of perfectly nested loops containing array references and computation codes, and generates optimized communication. We measured the performance of the output programs of the MCAMP optimizer by executing them on a real massively parallel machine GRAPE-DR using a software tool chain that we also implemented. We showed that MCAMP can achieve optimal data transfer patterns and comparable performance to that of hand-optimized codes with a short analysis time.","PeriodicalId":186773,"journal":{"name":"2008 International Conference on Parallel Architectures and Compilation Techniques (PACT)","volume":"55 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-10-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115833236","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}
Hyunchul Park, Kevin Fan, S. Mahlke, Taewook Oh, Heeseok Kim, Hong-Seok Kim
Coarse-grained reconfigurable architectures (CGRAs) present an appealing hardware platform by providing the potential for high computation throughput, scalability, low cost, and energy efficiency. CGRAs consist of an array of function units and register files often organized as a two dimensional grid. The most difficult challenge in deploying CGRAs is compiler scheduling technology that can efficiently map software implementations of compute intensive loops onto the array. Traditional schedulers focus on the placement of operations in time and space. With CGRAs, the challenge of placement is compounded by the need to explicitly route operands from producers to consumers. To systematically attack this problem, we take an edge-centric approach to modulo scheduling that focuses on the routing problem as its primary objective. With edge-centric modulo scheduling (EMS), placement is a by-product of the routing process, and the schedule is developed by routing each edge in the dataflow graph. Routing cost metrics provide the scheduler with a global perspective to guide selection. Experiments on a wide variety of compute-intensive loops from the multimedia domain show that EMS improves throughput by 25% over traditional iterative modulo scheduling, and achieves 98% of the throughput of simulated annealing techniques at a fraction of the compilation time.
{"title":"Edge-centric modulo scheduling for coarse-grained reconfigurable architectures","authors":"Hyunchul Park, Kevin Fan, S. Mahlke, Taewook Oh, Heeseok Kim, Hong-Seok Kim","doi":"10.1145/1454115.1454140","DOIUrl":"https://doi.org/10.1145/1454115.1454140","url":null,"abstract":"Coarse-grained reconfigurable architectures (CGRAs) present an appealing hardware platform by providing the potential for high computation throughput, scalability, low cost, and energy efficiency. CGRAs consist of an array of function units and register files often organized as a two dimensional grid. The most difficult challenge in deploying CGRAs is compiler scheduling technology that can efficiently map software implementations of compute intensive loops onto the array. Traditional schedulers focus on the placement of operations in time and space. With CGRAs, the challenge of placement is compounded by the need to explicitly route operands from producers to consumers. To systematically attack this problem, we take an edge-centric approach to modulo scheduling that focuses on the routing problem as its primary objective. With edge-centric modulo scheduling (EMS), placement is a by-product of the routing process, and the schedule is developed by routing each edge in the dataflow graph. Routing cost metrics provide the scheduler with a global perspective to guide selection. Experiments on a wide variety of compute-intensive loops from the multimedia domain show that EMS improves throughput by 25% over traditional iterative modulo scheduling, and achieves 98% of the throughput of simulated annealing techniques at a fraction of the compilation time.","PeriodicalId":186773,"journal":{"name":"2008 International Conference on Parallel Architectures and Compilation Techniques (PACT)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-10-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128211249","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}
Divya Gulati, Changkyu Kim, S. Sethumadhavan, S. Keckler, D. Burger
While technology trends have ushered in the age of chip multiprocessors (CMP), a fundamental question is what size to make each core. Most current commercial designs are symmetric CMPs (SCMP) in which each core is identical and range from a simple RISC processor to a complex out-of-order x86 processor. Some researchers have proposed asymmetric CMPs (ACMP) consisting of multiple types of cores. While less of an issue for ACMPs, the fixed nature of both these architectures makes them vulnerable to mismatches between the granularity of the cores and the parallelism in the workload, which can cause inefficient execution. To remedy this weakness, recent research has proposed flexible-core CMPs (FCMP), which have the capability of aggregating multiple small processing cores to form larger logical processors. FCMPs introduce a new resource allocation and scheduling problem which must determine how many logical processors should be configured, how powerful each processor should be, and where/when each task should run. This paper introduces and motivates this problem, describes the challenges associated with it, and evaluates algorithms appropriate for multitasking on FCMPs. We also evaluate static-core CMPs of various configurations and compare them to FCMPs for various multitasking workloads.
{"title":"Multitasking workload scheduling on flexible-core chip multiprocessors","authors":"Divya Gulati, Changkyu Kim, S. Sethumadhavan, S. Keckler, D. Burger","doi":"10.1145/1454115.1454142","DOIUrl":"https://doi.org/10.1145/1454115.1454142","url":null,"abstract":"While technology trends have ushered in the age of chip multiprocessors (CMP), a fundamental question is what size to make each core. Most current commercial designs are symmetric CMPs (SCMP) in which each core is identical and range from a simple RISC processor to a complex out-of-order x86 processor. Some researchers have proposed asymmetric CMPs (ACMP) consisting of multiple types of cores. While less of an issue for ACMPs, the fixed nature of both these architectures makes them vulnerable to mismatches between the granularity of the cores and the parallelism in the workload, which can cause inefficient execution. To remedy this weakness, recent research has proposed flexible-core CMPs (FCMP), which have the capability of aggregating multiple small processing cores to form larger logical processors. FCMPs introduce a new resource allocation and scheduling problem which must determine how many logical processors should be configured, how powerful each processor should be, and where/when each task should run. This paper introduces and motivates this problem, describes the challenges associated with it, and evaluates algorithms appropriate for multitasking on FCMPs. We also evaluate static-core CMPs of various configurations and compare them to FCMPs for various multitasking workloads.","PeriodicalId":186773,"journal":{"name":"2008 International Conference on Parallel Architectures and Compilation Techniques (PACT)","volume":"54 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-10-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131033034","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}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：