Sycl 2020

Reyes, Ruyman; Brown, Gordon; Burns, Rod; Wong, M. W.

doi:10.1145/3388333.3388649

Cited by 20 publications

(4 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Any device in the system can be programmed using a single programming model thanks to SYCL. This is a compelling proposal for the performance portability initiative [94]. In heterogeneous systems, where a computer node may be made up of multicore CPUs, GPUs, FPGAs, and other problem-oriented devices like ASICs (Application-Specific Integrated Circuits), it adds the data parallel programming model to C++ programs.…”

Section: G Syclmentioning

confidence: 99%

MIMD Programs Execution Support on SIMD Machines: A Holistic Survey

Mustafa,

Alkhasawneh,

Obeidat

et al. 2024

IEEE Access

View full text Add to dashboard Cite

The Single Instruction Multiple Data (SIMD) architecture, supported by various highperformance computing platforms, efficiently utilizes data-level parallelism. The SIMD model is used in traditional CPUs, dedicated vector systems, and accelerators such as GPUs, vector extensions, and Xeon Phi. It provides performance throughput in computation-intensive and data-parallel applications. Despite the similarity of data-processing principles between these architectures, porting various programming models between the reviewed platforms is challenging. Furthermore, enhancing the programmability of these architectures is an important feature for utilizing their emerging computing power and simplifying programming complexity. This paper reviews the basic principles of optimization techniques to run asynchronous Multiple Instruction Multiple Data (MIMD) on SIMD accelerators. It also surveys several GPU programming paradigms and application programming interfaces (APIs) and classifies these frameworks into different groups based on their criteria. In addition, a review of studies that performed a comparison of the collaborative execution of GPUs with CPUs and Xeon Phi is presented in this paper. This study will be beneficial for developers and researchers in the field of computer architecture and parallel computing of intensive scientific applications, specifically for early-stage high-performance computing researchers, to obtain a brief overview of performance optimization opportunities as well as the challenges of existing SIMD platforms.

show abstract

Section: G Syclmentioning

confidence: 99%

MIMD Programs Execution Support on SIMD Machines: A Holistic Survey

Mustafa,

Alkhasawneh,

Obeidat

et al. 2024

IEEE Access

View full text Add to dashboard Cite

show abstract

“…The first one, tested in [11,20], consists of running this code that provides the tasks and dependencies in a single process, which acts as master or client, while all the other processes involved act as slaves or servers, executing the tasks under its control. SYCL [18], which can manage multiple distinct memories and devices under this model from a single host application can also be classified in this group. The second strategy, followed by UPC++ DepSpawn and the MPI-based StarPU [2,3], executes the code in parallel in all the processes involved, so that they all have a view of the task graph and they agree on which process runs what and when.…”

Section: Related Workmentioning

confidence: 99%

The New UPC++ DepSpawn High Performance Library for Data-Flow Computing with Hybrid Parallelism

Fraguela

Andrade

2022

Lecture Notes in Computer Science

View full text Add to dashboard Cite

B asili o B. Fr a g u el a ( ) [ 0 0 0 0− 0 0 0 2 − 3 4 3 8 − 5 9 6 0] a n d Di e g o A n dr a d e [ 0 0 0 0− 0 0 0 1 − 5 6 7 0 − 7 4 2 5] U ni v e r si d a d e d a C o r u ñ a, CI TI C, G r u p o d e A r q uit e ct u r a d e C o m p ut a d o r e s. F a c ult a d e d e I nf o r m áti c a, C a m p u s d e El vi ñ a, S / N. 1 5 0 7 1. A C o r u ñ a, S p ai n. { b a s i l i o . f r a g u e l a , d i e g o . a n d r a d e } @ u d c . e s

show abstract

“…The need for a single application programming interface (API) to program their divergent hardware products -the currently dominant Xeon multi-core CPUs, recently announced Xe GPUs and Intel's FPGA devices -is driving this development. OneAPI is based on SYCL [2], a C++ abstraction layer for programming parallel systems, initially based on OpenCL, but as of the latest version of the standard, decoupled from it [20] to allow for different backends (e.g. CUDA, OpenMP).…”

Section: Introductionmentioning

confidence: 99%