检索结果-内蒙古大学图书馆

International Workshop on Performance, Portability and Productivity in HPC (P3HPC)

作者： Deakin, Tom McIntosh-Smith, Simon Pennycook, S. John Sewall, Jason Univ Bristol Dept Comp Sci Bristol Avon England Intel Corp Santa Clara CA USA

ISBN: (纸本)9781665424394

Reductions are a common pattern in parallel programming, and every parallel programming language or framework has its own reduction abstraction with its own idiosyncrasies. These abstractions differ not only in their syntax, but also in their semantics and their ability to express certain types of reduction. Such differences may prevent specific combinations of abstraction and hardware platform from reaching high levels of performance, with consequences for portability and programmer productivity. In this paper, we present a set of representative reduction benchmarks to explore the capabilities of five contemporary programming languages and frameworks - OpenMP, Kokkos, RAJA, SYCL, and the oneAPI DPC++ Library (oneDPL) - across a variety of hardware platforms, including CPUs and GPUs from multiple vendors. We discuss the advantages and disadvantages of each reduction abstraction, and conclude with recommendations to improve their design and implementation.

关键词： Productivity parallel programming Semantics Benchmark testing Writing Syntactics Hardware

来源：评论

学校读者我要写书评

暂无评论

A Dense Tensor Accelerator with Data Exchange Mesh for DNN and Vision Workloads 53

A Dense Tensor Accelerator with Data Exchange Mesh for DNN a...

引用

IEEE International Symposium on Circuits and Systems (IEEE ISCAS)

作者： Lin, Yu-Sheng Chen, Wei-Chao Yang, Chia-Lin Chien, Shao-Yi Inventec Corp Taipei Taiwan Natl Taiwan Univ Taipei Taiwan

ISBN: (纸本)9781728192017

We propose a dense tensor accelerator called VectorMesh, a scalable, memory-efficient architecture that can support a wide variety of DNN and computer vision workloads. Its building block is a tile execution unit (TEU), which includes dozens of processing elements (PEs) and SRAM buffers connected through a butterfly network. A mesh of FIFOs between the TEUs facilitates data exchange between tiles and promote local data to global visibility. Our design performs better according to the roofline model for CNN, GEMM, and spatial matching algorithms compared to state-of-the-art architectures. It can reduce global buffer and DRAM fetches by 2-22 times and up to 5 times, respectively.

关键词： Neural network hardware vector processors parallel programming

来源：评论

学校读者我要写书评

暂无评论

Concurrent and Distributed Pseudocode: A Systematic Literature Review 47

Concurrent and Distributed Pseudocode: A Systematic Literatu...

引用

47th Latin American Computing Conference (CLEI)

作者： Ulate-Caballero, Bryan Alexander Berrocal-Rojas, Allan Hidalgo-Cespedes, Jeisson Univ Costa Rica Comp & Informat San Jose Costa Rica Univ Costa Rica ECCI San Jose Costa Rica Univ Costa Rica ECCI CITIC San Jose Costa Rica

ISBN: (纸本)9781665495035

Pseudocode is a valuable resource used in programming education, software development, and scientific reports for designing algorithmic solutions as it is easy to write, understand, and modify. Since pseudocode is lacking in its ability to be tested, it is difficult to determine whether a pseudocode solution is correct or not. Software tools are specially required to reach this goal, e.g., helping professors find race conditions, deadlocks, or starvation issues while grading students' concurrent pseudocode. Although there are various tools to work with sequential pseudocode, there is a lack of tools to work with concurrent pseudocode. This shortage motivated us to determine the state-of-the-art in notations and tools for testing concurrent and distributed pseudocode. We conducted a systematic literature review and found only a few related publications, confirming that this topic is understudied. We found and report about five software tools capable of interpreting concurrent or distributed pseudocode, and two software tools capable of verifying its correctness. As another result, no other literature review was found about this topic, conferring novelty to the contributions of this work.

关键词： Concurrent pseudocode distributed pseudocode parallel pseudocode code correctness concurrent programming distributed programming parallel programming pseudocode pseudo-code pseudo-language education teaching

来源：评论

学校读者我要写书评

暂无评论

The parallel optimization based on the PVS algorithm and research on the evaluation function in the Game of the Amazons 33

The parallel optimization based on the PVS algorithm and res...

引用

33rd Chinese Control and Decision Conference (CCDC)

作者： Wang, Haoyu Qiu, Hongkun Shenyang Aerosp Univ Sch Comp Sci Shenyang 110136 Peoples R China Shenyang Aerosp Univ Engn Training Ctr Shenyang 110136 Peoples R China

ISBN: (纸本)9781665440899

The PVS search function, as a current mainstream and efficient algorithm, has been widely used in various kinds of chess program. We applied the parallel search function based on the PVS and improved the running speed of the program. At the same time, we also did some research and experiments on the evaluation function of Amazon chess which provided a set of available Amazon evaluation functions and parameter adjustment results for reference.

关键词： Computer Game Amazon Game PVS parallel programming Evaluation Function

来源：评论

学校读者我要写书评

暂无评论

Accelerating Messages by Avoiding Copies in an Asynchronous Task-based programming Model 6

Accelerating Messages by Avoiding Copies in an Asynchronous ...

引用

IEEE/ACM 6th International Workshop on Extreme Scale programming Models and Middleware (ESPM2)

作者： Bhat, Nitin White, Sam Ramos, Evan Kale, Laxmikant, V Charmworks Inc Urbana IL 61801 USA Univ Illinois Dept Comp Sci Urbana IL USA

ISBN: (纸本)9781665411400

Task-based programming models promise improved communication performance for irregular, fine-grained, and load imbalanced applications. They do so by relaxing some of the messaging semantics of stricter models and taking advantage of those at the lower-levels of the software stack. For example, while MPI's two-sided communication model guarantees in-order delivery, requires matching sends to receives, and has the user schedule communication, task-based models generally favor the runtime system scheduling all execution based on the dependencies and message deliveries as they happen. The messaging semantics are critical to enabling high performance. In this paper, we build on previous work that added zero copy semantics to Converse/LRTS. We examine the messaging semantics of Charm++ as it relates to large message buffers, identify shortcomings, and define new communication APIs to address them. Our work enables in-place communication semantics in the context of point-to-point messaging, broadcasts, transmission of read-only variables at program startup, and for migration of chares. We showcase the performance of our new communication APIs using benchmarks for Charm++ and Adaptive MPI, which result in nearly 90% latency improvement and 2x lower peak memory usage.

关键词： Charm plus AMPI RDMA parallel programming Asynchronous Tasking Communication Optimizations

来源：评论

学校读者我要写书评

暂无评论

Real-Time Scheduling Models in Diverse Multi-core OpenMP Applications

Real-Time Scheduling Models in Diverse Multi-core OpenMP App...

引用

International Conference on Decision Aid Sciences and Application (DASA)

作者： Waheed, Musfirah Nadeem Siddique, Mohammed Majan Univ Coll Fac Informat Technol Muscat Oman

ISBN: (纸本)9781665416344

This research presents some of the critical information required to understand the concept of parallel programming and the implementation of OpenMP in parallel programming. parallelism is the preferred tool for expediting an algorithm, as demonstrated by the evolution of computing architectures (multi-core and many-core) towards a greater number of processing cores. The report will focus on OpenMP parallel programming models and further examine its implementation and features. parallel programming OpenMP model is increasingly preferred for its ability to deliver real-time processing, thereby, meeting system requirements performance wise. Furthermore, the study of implementing OpenMP in enhancing the efficiency of 3D discontinuous deformation analysis (3D-DDA) for expansive simulation using parallel block Jacobi (BJ) and Pre-conditioned conjugate gradient (PCG) algorithms. The absence of synchronization of data in parallel programming makes the system more prone to errors in programming since the parallel environment is much more complicated than perceived. The studies performed will highlight how synchronization is managed using OpenMP model. In the field of biometrics, the most important issue faced in DNA sequencing and pattern discovery is locating the longest common subsequence (LCS) among sequences. To identify the LCS of DNA sequences, we will look into the solutions achieved using OpenMP mols based on CPU, that extend major improvements hi processing speed, capital, and ubiquity, and the results based on the analysis are discussed.

关键词： parallel programming openmp synchronization sporadic dimensional discontinuous deformation thread

来源：评论

学校读者我要写书评

暂无评论

parallel-batched Interpolation Search Tree

arXiv

引用

arXiv 2023年

作者： Aksenov, Vitaly Kokorin, Ilya Martsenyuk, Alena ITMO University Russia

A sorted set (or map) is one of the most used data types in computer science. In addition to standard set operations, like Insert, Remove, and Contains, it can provide set-set operations such as Union, Intersection, and Difference. Each of these set-set operations is equivalent to some batched operation: the data structure should be able to execute Insert, Remove, and Contains on a batch of keys. It is obvious that we want these "large" operations to be parallelized. These sets are usually implemented with the trees of logarithmic height, such as 2-3 trees, treaps, AVL trees, red-black trees, etc. Until now, little attention was devoted to data structures that work asymptotically better under several restrictions on the stored data. In this work, we parallelize Interpolation Search Tree which is expected to serve requests from a smooth distribution in doubly-logarithmic time. Our data structure of size n performs a batch of m operations in O(m log log n) work and poly-log span. © 2023, CC BY.

关键词： parallel programming

来源：评论

学校读者我要写书评

暂无评论

Combining One-Sided Communications with Task-Based programming Models

Combining One-Sided Communications with Task-Based Programmi...

引用

IEEE International Conference on Cluster Computing (Cluster)

作者： Sala, Kevin Macia, Sandra Beltran, Vicenc Barcelona Supercomp Ctr BSC Barcelona Spain

ISBN: (纸本)9781728196664

Hybrid programming combining task-based and message-passing models is an increasingly popular technique to exploit multi-core clusters. The Task-Aware MPI (TAMPI) library integrates both models enabling the safe overlap of computation and communication tasks using two-sided MPI communications. Two-sided primitives combine data transfers with implicit synchronizations, but one-sided models usually offer more efficient data transfers decoupling synchronizations. MPI offers four distinct one-sided synchronization modes, while GASPI is a PGAS API providing one-sided operations with remote notifications for fine inter-process synchronizations. In this paper, we study the challenges of integrating MPI and GASPI one-sided operations with the OpenMP and OmpSs-2 tasking models. We propose and implement several extensions to the GASPI and OmpSs-2 programming models, which are leveraged by a new library called Task-Aware GASPI (TAGASPI). The TAGASPI library allows the efficient and safe use of one-sided operations with remote notifications inside tasks. Both TAGASPI and TAMPI transparently manage communications issued by tasks and allow these to overlap with computation tasks naturally, following a data-flow model. These libraries are complementary and can be mixed in the same application. Our experience porting several mini-apps to this hybrid model shows that TAGASPI helps leverage one-sided communications with similar complexity to pure and hybrid two-sided MPI approaches. We show that our hybrid one-sided approach outperforms the pure MPI strategies, but it also surpasses the TAMPI's performance when stressing communication phases, e.g., increasing the communication parallelism and reducing the communication tasks' sizes.

关键词： One-sided RMA Communication parallel programming Message passing MPI GASPI Data-flow Task OpenMP OmpSs-2

来源：评论

学校读者我要写书评

暂无评论

Formal Verification of Distributed Master-Slave Finite State Machine 29

Formal Verification of Distributed Master-Slave Finite State...

引用

29th Telecommunications Forum (TELFOR)

作者： Popovic, Marko Marinkovic, Vladimir Djukic, Miodrag Popovic, Miroslav RT RK Inst Comp Based Syst Novi Sad Serbia Univ Novi Sad Fac Tech Sci Novi Sad Serbia

ISBN: (纸本)9781665425858

Recently, the distributed master-slave finite state machine appeared as a solution to govern a pair of transaction coordinators operating in the master-slave mode. In this paper, we formally verify the correctness of that solution using process algebra CSP and the model checker PAT. The CSP model consists of CSP processes modelling the transaction coordinators and individual master-slave finite state machines. During verification, altogether 30 assertions were automatically checked by PAT and found to be valid.

关键词： distributed systems parallel programming Python transactional memories fault tolerance formal verification

来源：评论

学校读者我要写书评

暂无评论

Distributed algorithms for scalable proximity operator computation and application to video denoising

引用

DIGITAL SIGNAL PROCESSING 2022年第0期128卷

作者： Abboud, Feriel Stamm, Marin Chouzenoux, Emilie Pesquet, Jean-Christophe Talbot, Hugues WITBE Collines Arche Immeuble Opera F-92800 Puteaux La Defense France Inst Polytech Paris ENSTA Paris Palaiseau France Univ Paris Saclay Ctr Visual Comp Cent Supelec Inria Gif Sur Yvette France

Optimization problems arising in signal and image processing involve an increasingly large number of variables. In addition to the curse of dimensionality, another difficulty to overcome is that the cost function usually reads as the sum of several loss or regularization terms, which are non-necessarily smooth and possibly composed with large-size linear operators. Proximal splitting approaches are fundamental tools to address such problems, with demonstrated efficiency in many applicative fields. In this paper, we present a new distributed algorithm for computing the proximity operator of a sum of non-necessarily smooth convex functions composed with arbitrary linear operators. Our algorithm relies on a primal-dual splitting strategy, and benefits from established convergence guaranties. Each involved function is associated with a node of a hypergraph, with the ability to communicate with neighboring nodes sharing the same hyperedge. Thanks to this structure, our method can be efficiently implemented on modern parallel computing architectures, distributing the computations on multiple nodes or machines, with controlled requirements for synchronization steps. Good numerical performance and scalability properties are demonstrated on a problem of video sequence denoising. Our code implemented in Julia is made available at https://github .com /MarinENSTA /distributed_julia_denoising. (C) 2022 Elsevier Inc. All rights reserved.

关键词： Convex optimization Proximal methods Distributed algorithm parallel programming Video restoration

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：