G. Balduzzi, Arghya Chatterjee, Ying Wai Li, P. Doak, U. Hähner, E. D'Azevedo, T. Maier, T. Schulthess
{"title":"Accelerating DCA++ (Dynamical Cluster Approximation) Scientific Application on the Summit Supercomputer","authors":"G. Balduzzi, Arghya Chatterjee, Ying Wai Li, P. Doak, U. Hähner, E. D'Azevedo, T. Maier, T. Schulthess","doi":"10.1109/PACT.2019.00041","DOIUrl":"https://doi.org/10.1109/PACT.2019.00041","url":null,"abstract":"Optimizing scientific applications on today's accelerator-based high performance computing systems can be challenging, especially when multiple GPUs and CPUs with heterogeneous memories and persistent non-volatile memories are present. An example is Summit, an accelerator-based system at the Oak Ridge Leadership Computing Facility (OLCF) that is rated as the world's fastest supercomputer to-date. New strategies are thus needed to expose the parallelism in legacy applications, while being amenable to efficient mapping to the underlying architecture. In this paper we discuss our experiences and strategies to port a scientific application, DCA++, to Summit. DCA++ is a high-performance research application that solves quantum many-body problems with a cutting edge quantum cluster algorithm, the dynamical cluster approximation. Our strategies aim to synergize the strengths of the different programming models in the code. These include: a) streamlining the interactions between the CPU threads and the GPUs, b) implementing computing kernels on the GPUs and decreasing CPU-GPU memory transfers, c) allowing asynchronous GPU communications, and d) increasing compute intensity by combining linear algebraic operations. Full-scale production runs using all 4600 Summit nodes attained a peak performance of 73.5 PFLOPS with a mixed precision implementation. We observed a perfect strong and weak scaling for the quantum Monte Carlo solver in DCA++, while encountering about 2x input/output (I/O) and MPI communication overhead on the time-to-solution for the full machine run. Our hardware agnostic optimizations are designed to alleviate the communication and I/O challenges observed, while improving the compute intensity and obtaining optimal performance on a complex, hybrid architecture like Summit.","PeriodicalId":444467,"journal":{"name":"2019 28th International Conference on Parallel Architectures and Compilation Techniques (PACT)","volume":"37 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-11-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130335946","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":"Absinthe: Learning an Analytical Performance Model to Fuse and Tile Stencil Codes in One Shot","authors":"Tobias Gysi, T. Grosser, T. Hoefler","doi":"10.1109/PACT.2019.00036","DOIUrl":"https://doi.org/10.1109/PACT.2019.00036","url":null,"abstract":"Expensive data movement makes the optimal target-specific selection of data-locality transformations essential. Loop fusion and tiling are the most important data-locality transformations. Their optimal selection is hard since good tile size choices inherently depend on the fusion choices and vice versa. Existing approaches avoid this difficulty by optimizing independent analytical models or by reverting to heuristics. Absinthe formulates the first unified linear optimization problem to derive single shot fusion and tile size decisions for stencil codes. At the core of our optimization problem, we place a learned analytic performance model that captures the characteristics of the target system. The tuned application kernels demonstrate excellent performance within 10% of exhaustively auto-tuned versions and up to 74% faster than the results of independent optimization with max fusion heuristic and Absinthe tile size selection. While the full search space is non-linear, bounding it to relevant solutions enables the efficient exploration of the exponential search space using linear solvers. As a result, the tuning of our application kernels takes less than one minute. Our approach thus establishes the foundations for next-generation compilers, which exploit empirical information to guide target-specific code transformations.","PeriodicalId":444467,"journal":{"name":"2019 28th International Conference on Parallel Architectures and Compilation Techniques (PACT)","volume":"54 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-11-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131368340","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}
Teng Yu, Pavlos Petoumenos, Vladimir Janjic, Mingcan Zhu, Hugh Leather, John Thomson
{"title":"POSTER: A Collaborative Multi-Factor Scheduler for Asymmetric Multicore Processors","authors":"Teng Yu, Pavlos Petoumenos, Vladimir Janjic, Mingcan Zhu, Hugh Leather, John Thomson","doi":"10.1109/PACT.2019.00058","DOIUrl":"https://doi.org/10.1109/PACT.2019.00058","url":null,"abstract":"Asymmetric multicore processors (AMP) are necessary for extracting performance in an era of limited power budget and dark silicon. We have efficient symmetric schedulers, efficient asymmetric schedulers for single-threaded workloads, and efficient asymmetric schedulers for single program workloads. What we do not have is a scheduler that can handle all three factors affecting AMP scheduling: core affinity, thread criticality, and scheduling fairness. To address this problem, this paper introduces the first general purpose asymmetry-aware scheduler targeting multi-threaded multi-programmed workloads. It estimates the performance of each thread on each type of core and it identifies communication patterns and bottleneck threads. With this information, the scheduler makes coordinated core assignment and thread selection decisions that still provide each application its fair share of the processor's time. We evaluated our approach on GEM5 through four distinct big.LITTLE configurations and multi-threaded multi-programmed workloads composed of PARSEC and SPLASH2 benchmarks. Compared against the Linux CFS scheduler and a state-of-the-art AMP-aware scheduler, we demonstrate performance gains of up to 25% and 5% to 15% on average depending on the hardware setup.","PeriodicalId":444467,"journal":{"name":"2019 28th International Conference on Parallel Architectures and Compilation Techniques (PACT)","volume":"62 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-10-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116337676","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}
Jongwook Chung, Yuhwan Ro, Joonsung Kim, Jaehyung Ahn, Jangwoo Kim, John Kim, Jae W. Lee, Jung Ho Ahn
{"title":"Enforcing Last-Level Cache Partitioning through Memory Virtual Channels","authors":"Jongwook Chung, Yuhwan Ro, Joonsung Kim, Jaehyung Ahn, Jangwoo Kim, John Kim, Jae W. Lee, Jung Ho Ahn","doi":"10.1109/PACT.2019.00016","DOIUrl":"https://doi.org/10.1109/PACT.2019.00016","url":null,"abstract":"Ensuring fairness or providing isolation between multiple workloads with different characteristics that are colocated on a single, shared-memory system is a challenge. Recent multicore processors provide last-level cache (LLC) hardware partitioning to provide hardware support for isolation, with the cache partitioning often specified by the user. While more LLC capacity often results in higher performance, in this work we identify that a workload allocated more LLC capacity result in worse performance on real-machine experiments, which we refer to as MiW (more is worse). Through various controlled experiments, we identify that another workload with less LLC capacity causes more frequent LLC misses. The workload stresses the main-memory system shared by both workloads and degrades the performance of the former workload even if the LLC partitioning is used (a balloon effect). To resolve this problem, we propose virtualizing the datapath of main-memory controllers and dedicating the memory virtual channels (mVCs) to each group of applications, grouped for LLC partitioning. mVC can further fine-tune the performance of groups by differentiating buffer sizes among mVCs. It can reduce the total system cost by executing latency-critical and throughput-oriented workloads together on shared machines, of which performance criteria can be achieved only on dedicated machines if mVCs are not supported. Experiments on a simulated chip multiprocessor show that our proposals effectively eliminate the MiW phenomenon, hence providing additional opportunities for workload consolidation in a datacenter. Our case study demonstrates potential savings of machine count by 21.8% with mVC, which would otherwise violate a service level objective (SLO).","PeriodicalId":444467,"journal":{"name":"2019 28th International Conference on Parallel Architectures and Compilation Techniques (PACT)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-09-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130819250","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}
Adrián Barredo, J. M. Cebrian, Miquel Moretó, Marc Casas, M. Valero
{"title":"POSTER: An Optimized Predication Execution for SIMD Extensions","authors":"Adrián Barredo, J. M. Cebrian, Miquel Moretó, Marc Casas, M. Valero","doi":"10.1109/PACT.2019.00054","DOIUrl":"https://doi.org/10.1109/PACT.2019.00054","url":null,"abstract":"Vector processing is a widely used technique to improve performance and energy efficiency in modern processors. Most of them rely on predication to support divergence control. However, performance and energy consumption in predicated instructions are usually independent on the number of true values in a mask. This means that the efficiency of the system becomes sub-optimal as vector length increases. In this work we propose the Optimized Predication Execution (OPE) technique. OPE delays the execution of sparse masked vector instructions sharing the same PC, extracts their active elements and creates a new dense instruction with a higher mask density. After executing such dense instruction, results are restored to the original sparse instructions. Our approach improves performance by up to 25% and reduces dynamic energy consumption by up to 43% on real applications with predication.","PeriodicalId":444467,"journal":{"name":"2019 28th International Conference on Parallel Architectures and Compilation Techniques (PACT)","volume":"57 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126813642","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":"POSTER: Leveraging Run-Time Feedback for Efficient ASR Acceleration","authors":"R. Yazdani, J. Arnau, Antonio González","doi":"10.1109/PACT.2019.00046","DOIUrl":"https://doi.org/10.1109/PACT.2019.00046","url":null,"abstract":"In this work, we propose Locality-AWare-Scheme (LAWS) for an Automatic Speech Recognition (ASR) accelerator in order to significantly reduce its energy consumption and memory requirements, by leveraging the locality among consecutive segments of the speech signal. LAWS diminishes ASR's workload by up to 60% by removing most of the off-chip accesses during the ASR's decoding process. We furthermore improve LAWS's effectiveness by selectively adapting the amount of ASR's workload, based on run-time feedback. In particular, we exploit the fact that the confidence of the ASR system varies along the recognition process. When confidence is high, the ASR system can be more restrictive and reduce the amount of work. The end design provides a saving of 87% in memory requests, 2.3x reduction in energy consumption, and a speedup of 2.1x with respect to the state-of-the-art ASR accelerator.","PeriodicalId":444467,"journal":{"name":"2019 28th International Conference on Parallel Architectures and Compilation Techniques (PACT)","volume":"50 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122863449","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":"POSTER: Variable Sized Cache-Block Compaction","authors":"Sayantan Ray, M. Mutyam","doi":"10.1109/PACT.2019.00050","DOIUrl":"https://doi.org/10.1109/PACT.2019.00050","url":null,"abstract":"Data blocks compressed to different sizes can be stored together inside a single cache-block to increase space utilization. However, the lack of a common size offset makes it challenging to locate individual blocks without additional tag overhead. We propose Variable Sized Cache-Block Compaction (VSCC) that allows us to store variable sized compressed blocks together and locate them inside a cache-block by using their compression encodings – available inside the tag metadata. We introduce a novel read/write scheme and a new BDI compression encoding, which reduce the necessary operations by 50%. Experimental results reveal that VSCC outperforms state-of-the-art techniques from the performance and energy point of view while keeping the storage overheads within acceptable limits.","PeriodicalId":444467,"journal":{"name":"2019 28th International Conference on Parallel Architectures and Compilation Techniques (PACT)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123021487","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":"POSTER: SPiDRE: Accelerating Sparse Memory Access Patterns","authors":"Adrián Barredo, J. Beard, Miquel Moretó","doi":"10.1109/PACT.2019.00056","DOIUrl":"https://doi.org/10.1109/PACT.2019.00056","url":null,"abstract":"Development in process technology has led to an exponential increase in processor speed and memory capacity. However, memory latencies have not improved as dramatically and represent a well-known problem in computer architecture. Cache memories provide more bandwidth with lower latencies than main memories but they are capacity limited. Locality-friendly applications benefit from a large and deep cache hierarchy. Nevertheless, this is a limited solution for applications suffering from sparse and irregular memory access patterns, such as data analytics. In order to accelerate them, we should maximize usable bandwidth, reduce latency and maximize moved data reuse. In this work we explore the Sparse Data Rearrange Engine (SPiDRE), a novel hardware approach to accelerate these applications through near-memory data reorganization.","PeriodicalId":444467,"journal":{"name":"2019 28th International Conference on Parallel Architectures and Compilation Techniques (PACT)","volume":"61 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116713638","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":"Deepframe: A Profile-Driven Compiler for Spatial Hardware Accelerators","authors":"Apala Guha, Naveen Vedula, Arrvindh Shriraman","doi":"10.1109/PACT.2019.00014","DOIUrl":"https://doi.org/10.1109/PACT.2019.00014","url":null,"abstract":"Tracing code paths to form extended basic blocks is useful in many areas, compiler optimizations [1], improving instruction cache behavior [2] and custom-hardware offloading [3]. Prior work has been plagued by small traces, limited either by the overheads of dynamic profiling, statically available information [4], or side-exit branches [5]. In this work, we rethink what code path sequences to fuse and construct long traces for offloading to spatial accelerators, while minimizing the occurrence of side exits which limit dynamic coverage. We introduce a novel technique that recasts learning a program's execution patterns as a natural-language-processing problem, CBOW (Continuous Bag of Words). We then use a deep learning network to learn the relationships among paths. During the compilation phase, the compiler uses a sequence miner to decide what paths are likely to occur. The learning network predicts a Deepframe online, which is an extended basic block comprising a multi-path sequence (each path itself is composed of multiple basic blocks). We demonstrate the efficacy of Deepframe on spatial hardware accelerators and find the following: i) Deepframe can construct up to 5x (max: 27x) longer offload regions compared to prior approaches. ii) Surprisingly far-flung ILP (instruction-level parallelism) and MLP (memory-level parallelism) can be mined from the frames statically (5.5x increase in ILP and 10.5x increase in MLP). iii) The frames offloaded to the spatial accelerator have minimal side exits (mis-speculation) and achieve sufficient dynamic coverage to improve overall application performance (up to 9x improvement). We will be releasing open-source our end-to-end compiler prototype based on LLVM.","PeriodicalId":444467,"journal":{"name":"2019 28th International Conference on Parallel Architectures and Compilation Techniques (PACT)","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131573402","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}
Yuan Wen, Andrew Anderson, Valentin Radu, M. O’Boyle, David Gregg
{"title":"POSTER: Space and Time Optimal DNN Primitive Selection with Integer Linear Programming","authors":"Yuan Wen, Andrew Anderson, Valentin Radu, M. O’Boyle, David Gregg","doi":"10.1109/PACT.2019.00059","DOIUrl":"https://doi.org/10.1109/PACT.2019.00059","url":null,"abstract":"Convolutional neural networks (CNNs) are used in many applications, from industrial robotics to biometric identification on mobile devices. But they can be too resource-hungry for mobile and embedded devices with tightly constrained memory and energy budgets. We propose an ahead-of-time primitive selection for CNNs, based on integer linear programming (ILP). Under a tight memory budget, our ILP solver selects the optimal primitive for each layer such that the entire network is optimized for execution time subject to a memory budget, or vice versa. Our method yields significant speedup and memory reduction compared to existing methods.","PeriodicalId":444467,"journal":{"name":"2019 28th International Conference on Parallel Architectures and Compilation Techniques (PACT)","volume":"12 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131082793","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}