检索结果-内蒙古大学图书馆

proceedings of the 16th ACM symposium on principles and practice of parallel programming

作者： Kathryn S. McKinley The University of Texas at Austin Austin TX USA

ISBN: (纸本)9781450301190

Two trends changed the computing landscape over the past decade: (1) hardware vendors started delivering chip multiprocessors (CMPs) instead of uniprocessors, and (2) software developers increasingly chose managed languages instead of native languages. Unfortunately, the former change is disrupting the virtuous-cycle between performance improvements and software innovation. Establishing a new parallel performance virtuous cycle for managed languages will require scalable applications executing on scalable Virtual Machine (VM) services, since the VM schedules, monitors, compiles, optimizes, garbage collects, and executes together with the application. this talk describes current progress, opportunities, and challenges for scalable VM services. the parallel computing revolution urgently needs more innovations.

关键词： multicore managed languages

来源：评论

学校读者我要写书评

暂无评论

Beyond parallel programming with domain specific languages 14

Beyond parallel programming with domain specific languages

引用

proceedings of the 19th ACM SIGPLAN symposium on principles and practice of parallel programming

作者： Kunle Olukotun Stanford University Stanford CA USA

ISBN: (纸本)9781450326568

Today, almost all computer architectures are parallel and heterogeneous; a combination of multiple CPUs, GPUs and specialized processors. this creates a challenging problem for application developers who want to develop high performance programs without the effort required to use low-level, architecture specific parallel programming models (e.g. OpenMP for CMPs, CUDA for GPUs, MPI for clusters). Domain-specific languages (DSLs) are a promising solution to this problem because they can provide an avenue for high-level application-specific abstractions with implicit parallelism to be mapped directly to low level architecture-specific programming models; providing both high programmer productivity and high execution *** this talk I will describe an approach to building high performance DSLs, which is based on DSL embedding in a general purpose programming language, metaprogramming and a DSL infrastructure called Delite. I will describe how we transform DSL programs into efficient first-order low-level code using domain specific optimization, parallelism and locality optimization with parallel patterns, and architecture-specific code generation. All optimizations and transformations are implemented in Delite: an extensible DSL compiler infrastucture that significantly reduces the effort required to develop new DSLs. Delite DSLs for machine learning, data querying, graph analysis, and scientific computing all achieve performance competitive with manually parallelized C++ code.

关键词： domain specific languages

来源：评论

学校读者我要写书评

暂无评论

Stream-K: Work-Centric parallel Decomposition for Dense Matrix-Matrix Multiplication on the GPU 23

Stream-K: Work-Centric Parallel Decomposition for Dense Matr...

引用

proceedings of the 28th ACM SIGPLAN Annual symposium on principles and practice of parallel programming

作者： Muhammad Osama Duane Merrill Cris Cecka Michael Garland John D. Owens University of California Davis NVIDIA Corporation

ISBN: (纸本)9798400700156

We introduce Stream-K, a work-centric parallelization of matrix multiplication (GEMM) and related computations in dense linear algebra. Whereas contemporary decompositions are primarily tile-based, our method operates by partitioning an even share of the aggregate inner loop iterations among physical processing elements. this provides a near-perfect utilization of computing resources, regardless of how efficiently the output tiling for any given problem quantizes across the underlying processing elements.

关键词： GPU matrix-multiplication load-balancing

来源：评论

学校读者我要写书评

暂无评论

Automatic static cost analysis for parallel programs

引用

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) 2015年 9032卷 132-157页

作者： Hoffmann, Jan Shao, Zhong Yale University United States

ISBN: (纸本)9783662466681

Static analysis of the evaluation cost of programs is an extensively studied problem that has many important applications. However, most automatic methods for static cost analysis are limited to sequential evaluation while programs are increasingly evaluated on modern multicore and multiprocessor hardware. this article introduces the first automatic analysis for deriving bounds on the worst-case evaluation cost of parallel first-order functional programs. the analysis is performed by a novel type system for amortized resource analysis. the main innovation is a technique that separates the reasoning about sizes of data structures and evaluation cost within the same framework. the cost semantics of parallel programs is based on call-by-value evaluation and the standard cost measures work and depth. A soundness proof of the type system establishes the correctness of the derived cost bounds with respect to the cost semantics. the derived bounds are multivariate resource polynomials which depend on the sizes of the arguments of a function. Type inference can be reduced to linear programming and is fully automatic. A prototype implementation of the analysis system has been developed to experimentally evaluate the effectiveness of the approach. the experiments show that the analysis infers bounds for realistic example programs such as quick sort for lists of lists, matrix multiplication, and an implementation of sets with lists. the derived bounds are often asymptotically tight and the constant factors are close to the optimal ones. © Springer-Verlag Berlin Heidelberg 2015.

关键词： Static analysis

来源：评论

学校读者我要写书评

暂无评论

parallel algorithms for masked sparse matrix-matrix products 22

Parallel algorithms for masked sparse matrix-matrix products

引用

proceedings of the 27th ACM SIGPLAN symposium on principles and practice of parallel programming

作者： Srđan Milaković Oguz Selvitopi Israt Nisa Zoran Budimlić Aydin Buluç Rice University Lawrence Berkeley Nat. Laboratory AWS AI

ISBN: (纸本)9781450392044

Computing the product of two sparse matrices (SpGEMM) is a fundamental operation in various combinatorial and graph algorithms as well as various bioinformatics and data analytics applications for computing inner-product similarities. For an important class of algorithms, only a subset of the output entries are needed, and the resulting operation is known as Masked SpGEMM since a subset of the output entries is considered to be "masked out". In this work, we investigate various novel algorithms and data structures for this rather challenging and important computation, and provide guidelines on how to design a fast Masked-SpGEMM for shared-memory architectures.

关键词： GraphBLAS

来源：评论

学校读者我要写书评

暂无评论

A parallel branch-and-bound algorithm with history-based domination 22

A parallel branch-and-bound algorithm with history-based dom...

引用

proceedings of the 27th ACM SIGPLAN symposium on principles and practice of parallel programming

作者： Taspon Gonggiatgul Ghassan Shobaki Pinar Muyan-Özçelik California State University

ISBN: (纸本)9781450392044

In this paper, we describe a parallel Branch-and-Bound (B&B) algorithm with a history-based domination technique, and we apply it to the Sequential Ordering Problem (SOP). To the best of our knowledge, the proposed algorithm is the first parallel B&B algorithm that includes a history-based domination technique and is the first parallel B&B algorithm for solving the SOP using a pure B&B approach. the proposed algorithm takes a pool-based approach and employs a collection of novel techniques that we have developed to achieve effective parallel exploration of the solution space, including parallel history domination, history table memory management, and a thread restart technique. the proposed algorithm was experimentally evaluated using the SOPLIB and TSPLIB benchmarks. the results show that using ten threads with a time limit of one hour on the medium-difficulty instances, the proposed algorithm gives a geometric-mean speedup of 19.9 on SOPLIB and 10.23 on TSPLIB, with super-linear speedups up to 65x seen on 17 instances.

关键词： NP-complete problems

来源：评论

学校读者我要写书评

暂无评论

Fast parallel Exact Inference on Bayesian Networks 23

Fast Parallel Exact Inference on Bayesian Networks

引用

proceedings of the 28th ACM SIGPLAN Annual symposium on principles and practice of parallel programming

作者： Jiantong Jiang Zeyi Wen Atif Mansoor Ajmal Mian The University of Western Australia Hong Kong University of Science and Technology (Guangzhou)

ISBN: (纸本)9798400700156

Bayesian networks (BNs) are attractive, because they are graphical and interpretable machine learning models. However, exact inference on BNs is time-consuming, especially for complex problems. To improve the efficiency, we propose a fast BN exact inference solution named Fast-BNI on multi-core CPUs. Fast-BNI enhances the efficiency of exact inference through hybrid parallelism that tightly integrates coarse- and fine-grained parallelism. We also propose techniques to further simplify the bottleneck operations of BN exact inference. Fast-BNI source code is freely available at https://***/jjiantong/FastBN.

关键词： junction tree inference bayesian networks

来源：评论

学校读者我要写书评

暂无评论

CuPBoP: A Framework to Make CUDA Portable 23

CuPBoP: A Framework to Make CUDA Portable

引用

proceedings of the 28th ACM SIGPLAN Annual symposium on principles and practice of parallel programming

作者： Ruobing Han Jun Chen Bhanu Garg Jeffrey Young Jaewoong Sim Hyesoon Kim Georgia Institute of Technology Seoul National University

ISBN: (纸本)9798400700156

CUDA, as one of the most popular choices for GPU programming, can be executed only on NVIDIA GPUs. To execute CUDA on non-NVIDIA devices, researchers have proposed to translate CUDA to other programming languages. However, this approach cannot achieve high coverage due to the challenges in source-to-source *** propose a framework, CuPBoP, that executes CUDA programs on non-NVIDIA devices without relying on other programming languages. CuPBoP consists of two parts. the compilation part applies transformations on CUDA host/kernel IRs. the runtime part consists of the runtime libraries for CUDA built-in functions. For the CPU backends, compared with the existing frameworks, CuPBoP achieves the highest coverage on all CPUs that we evaluate (x86, aarch64, RISC-V).We make CuPBoP publicly available to inspire more works in this area 1.

关键词： AArch64 X86 portable programming CUDA RISC-V

来源：评论

学校读者我要写书评

暂无评论

Code motion for explicitly parallel programs 99

Code motion for explicitly parallel programs

引用

proceedings of the seventh ACM SIGPLAN symposium on principles and practice of parallel programming

作者： Jens Knoop Bernhard Steffen Fachbereich Informatik Universität Dortmund Baroper Straβe 301 D-44227 Dortmund Germany

ISBN: (纸本)9781581131000

In comparison to automatic parallelization, which is thoroughly studied in the literature [31, 33], classical analyses and optimizations of explicitly parallel programs were more or less neglected. this may be due to the fact that naive adaptations of the sequential techniques fail [24], and their straightforward correct ones have unacceptable costs caused by the interleavings, which manifest the possible executions of a parallel program. Recently, however, we showed that unidirectional bitvector analyses can be performed for parallel programs as easily and as efficiently as for sequential ones [17], a necessary condition for the successful transfer of the classical optimizations to the parallel *** this article we focus on possible subsequent code motion transformations, which turn out to require much more care than originally conjectured [17]. Essentially, this is due to the fact that interleaving semantics, although being adequate for correctness considerations, fails when it comes to reasoning about efficiency of parallel programs. this deficiency, however, can be overcome by strengthening the specific treatment of synchronization points.

关键词： code motion (partial redundancy elimination) interleaving semantics shared memory bitvector problems synchronization code optimization data-flow analysis

来源：评论

学校读者我要写书评

暂无评论

LB-HM: load balance-aware data placement on heterogeneous memory for task-parallel HPC applications 22

LB-HM: load balance-aware data placement on heterogeneous me...

引用

proceedings of the 27th ACM SIGPLAN symposium on principles and practice of parallel programming

作者： Zhen Xie Jie Liu Sam Ma Jiajia Li Dong Li University of California College of William & Mary

ISBN: (纸本)9781450392044

the emergence of heterogeneous memory (HM) provides a cost-effective and high-performance solution to memory-consuming HPC applications. However, using HM, wisely migrating data objects on it is critical for high performance. In this work, we introduce a load balance-aware page management system, named LB-HM. LB-HM introduces task semantics during memory profiling, rather than being application-agnostic. Evaluating with a set of memory-consuming HPC applications, we show that we show that LB-HM reduces existing load imbalance and leads to an average of 17.1% and 15.4% (up to 26.0% and 23.2%) performance improvement, compared with a hardware-based solution and an industry-quality software-based solution on Optane-based HM.

关键词：

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：