Where is the data? Why you cannot debate CPU vs. GPU performance without the answer

Gregg, Chris; Hazelwood, Kim

doi:10.1109/ispass.2011.5762730

Cited by 209 publications

(146 citation statements)

References 18 publications

Supporting

Mentioning

145

Contrasting

Order By: Relevance

“…He and others observed that joins are 2-7 times faster on the GPU, whereas selections are 2-4 times slower, due to the required data transfers [30]. The same conclusion was made by Gregg and others, who showed that a GPU algorithm is not neccesarily faster than its CPU counterpart, due to the expensive data transfers [27]. One major point for achieving good performance in a GDBMS is therefore to avoid data transfers where possible.…”

Section: Non-functional Propertiesmentioning

confidence: 69%

GPU-Accelerated Database Systems: Survey and Open Challenges

Breß

Heimel

Siegmund

et al. 2014

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. The vast amount of processing power and memory bandwidth provided by modern graphics cards make them an interesting platform for data-intensive applications. Unsurprisingly, the database research community identified GPUs as effective co-processors for data processing several years ago. In the past years, there were many approaches to make use of GPUs at different levels of a database system. In this paper, we explore the design space of GPU-accelerated database management systems. Based on this survey, we present key properties, important trade-offs and typical challenges of GPU-aware database architectures, and identify major open challenges. Additionally, we survey existing GPU-accelerated DBMSs and classify their architectural properties. Then, we summarize typical optimizations implemented in GPU-accelerated DBMSs. Finally, we propose a reference architecture, indicating how GPU acceleration can be integrated in existing DBMSs.

show abstract

Section: Non-functional Propertiesmentioning

confidence: 69%

GPU-Accelerated Database Systems: Survey and Open Challenges

Breß

Heimel

Siegmund

et al. 2014

Lecture Notes in Computer Science

View full text Add to dashboard Cite

show abstract

“…One obvious benefit is avoiding data transfer through connection interfaces (e.g., PCIe link), which is one of the most well known bottlenecks of co-processor computing [34]. Additionally, GPU cores can access more memory by paging memory to and from disk.…”

Section: Heterogeneous Processorsmentioning

confidence: 99%

Speculative segmented sum for sparse matrix-vector multiplication on heterogeneous processors

Liu

Vinter

2015

Parallel Computing

View full text Add to dashboard Cite

Sparse matrix-vector multiplication (SpMV) is a central building block for scientific software and graph applications. Recently, heterogeneous processors composed of different types of cores attracted much attention because of their flexible core configuration and high energy efficiency. In this paper, we propose a compressed sparse row (CSR) format based SpMV algorithm utilizing both types of cores in a CPU-GPU heterogeneous processor. We first speculatively execute segmented sum operations on the GPU part of a heterogeneous processor and generate a possibly incorrect results. Then the CPU part of the same chip is triggered to re-arrange the predicted partial sums for a correct resulting vector. On three heterogeneous processors from Intel, AMD and nVidia, using 20 sparse matrices as a benchmark suite, the experimental results show that our method obtains significant performance improvement over the best existing CSR-based SpMV algorithms.

show abstract

“…Two work-items process one row of the block. The work-item with the even ID reads In[0] to In [4] to produce an eightpixel row from Out[0] to Out [7], and the work-item with the odd ID reads In [4] to In [7] to produce the successive eight-pixel row Out [8] to Out [15].…”

Section: Upsamplingmentioning

confidence: 99%

Dynamic partitioning‐based JPEG decompression on heterogeneous multicore architectures

Sodsong

Hong

Chung

et al. 2015

Concurrency and Computation

View full text Add to dashboard Cite

With the emergence of social networks and improvements in computational photography, billions of JPEG images are shared and viewed on a daily basis. Desktops, tablets and smartphones constitute the vast majority of hardware platforms used for displaying JPEG images. Despite the fact that these platforms are heterogeneous multicores, no approach exists yet that is capable of joining forces of a system's CPU and GPU for JPEG decoding.In this paper we introduce a novel JPEG decoding scheme for heterogeneous architectures consisting of a CPU and an OpenCLprogrammable GPU. We employ an offline profiling step to determine the performance of a system's CPU and GPU with respect to JPEG decoding. For a given JPEG image, our performance model uses (1) the CPU and GPU performance characteristics, (2) the image entropy and (3) the width and height of the image to balance the JPEG decoding workload on the underlying hardware. Our runtime partitioning and scheduling scheme exploits task, data and pipeline parallelism by scheduling the non-parallelizable entropy decoding task on the CPU, whereas inverse cosine transformations (IDCTs), color conversions and upsampling are conducted on both the CPU and the GPU. Our kernels have been optimized for GPU memory hierarchies.We have implemented the proposed method in the context of the libjpeg-turbo library, which is an industrial-strength JPEG encoding and decoding engine. Libjpeg-turbo's hand-optimized SIMD routines for ARM and x86 constitute a competitive yardstick for the comparison to the proposed approach. Retro-fitting our method with libjpeg-turbo provides insights on the software-engineering aspects of re-engineering legacy code for heterogeneous multicores.We have evaluated our approach for a total of 7194 JPEG images across three high-and middle-end CPU-GPU combinations. We achieve speedups of up to 4.2x over the SIMD-version of libjpeg-turbo, and speedups of up to 8.5x over its sequential code. Taking into account the non-parallelizable JPEG entropy decoding part, our approach achieves up to 95% of the theoretically attainable maximal speedup, with an average of 88%.

show abstract

Where is the data? Why you cannot debate CPU vs. GPU performance without the answer

Cited by 209 publications

References 18 publications

GPU-Accelerated Database Systems: Survey and Open Challenges

GPU-Accelerated Database Systems: Survey and Open Challenges

Speculative segmented sum for sparse matrix-vector multiplication on heterogeneous processors

Dynamic partitioning‐based JPEG decompression on heterogeneous multicore architectures

Contact Info

Product

Resources

About