2010arXiv (Cornell University)Open access

CrystalGPU: Transparent and Efficient Utilization of GPU Power

Abdullah Gharaibeh, Samer Al-Kiswany, Matei Ripeanu

Open full text 7 citations

Abstract

General-purpose computing on graphics processing units (GPGPU) has recently gained considerable attention in various domains such as bioinformatics, databases and distributed computing. GPGPU is based on using the GPU as a co-processor accelerator to offload computationally-intensive tasks from the CPU. This study starts from the observation that a number of GPU features (such as overlapping communication and computation, short lived buffer reuse, and harnessing multi-GPU systems) can be abstracted and reused across different GPGPU applications. This paper describes CrystalGPU, a modular framework that transparently enables applications to exploit a number of GPU optimizations. Our evaluation shows that CrystalGPU enables up to 16x speedup gains on synthetic benchmarks, while introducing negligible latency overhead.

Open-access reader

About this research paper

What this paper is about

General-purpose computing on graphics processing units (GPGPU) has recently gained considerable attention in various domains such as bioinformatics, databases and distributed computing. GPGPU is based on using the GPU as a co-processor accelerator to offload computationally-intensive tasks from the CPU. This study starts from the observation that a number of GPU features (such as overlapping communication and computation, short lived buffer reuse, and harnessing multi-GPU systems) can be abstracted and reused across different GPGPU applications. This paper describes CrystalGPU, a modular framework that transparently enables applications to exploit a number of GPU optimizations. Our evaluation shows that CrystalGPU enables up to 16x speedup gains on synthetic benchmarks, while introducing negligible latency overhead.

Why it matters

OpenAlex reports 7 citations for this work. Citation counts describe recorded attention and do not establish research quality.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

General-purpose computing on graphics processing units (GPGPU) has recently gained considerable attention in various domains such as bioinformatics, databases and distributed computing. GPGPU is based on using the GPU as a co-processor accelerator to offload computationally-intensive tasks from the CPU. This study starts from the observation that a number of GPU features (such as overlapping communication and computation, short lived buffer reuse, and harnessing multi-GPU systems) can be abstracted and reused across different GPGPU applications. This paper describes CrystalGPU, a modular framework that transparently enables applications to exploit a number of GPU optimizations. Our evaluation shows that CrystalGPU enables up to 16x speedup gains on synthetic benchmarks, while introducing negligible latency overhead.

Key concepts: Computer science, Parallel computing, General-purpose computing on graphics processing units, CUDA, Power (physics), Computational science, Computer graphics (images), Graphics

Related papers

Back to paper searchBrowse research topicsOriginal source
CrystalGPU: Transparent and Efficient Utilization of GPU Power — Research Paper | ScholarLens