pFPC: A Parallel Compressor for Floating-Point Data

Burtscher, Martin; Ratanaworabhan, Paruj

doi:10.1109/dcc.2009.43

Cited by 13 publications

(13 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The reported performance is for a single thread and can possibly be improved significantly with a parallel implementation, as we have done with FPC [20].…”

Section: Gfpc Throughputmentioning

confidence: 87%

gFPC: A Self-Tuning Compression Algorithm

Burtscher

Ratanaworabhan

2010

2010 Data Compression Conference

Self Cite

View full text Add to dashboard Cite

show abstract

“…The reported performance is for a single thread and can possibly be improved significantly with a parallel implementation, as we have done with FPC [20].…”

Section: Gfpc Throughputmentioning

confidence: 87%

gFPC: A Self-Tuning Compression Algorithm

Burtscher

Ratanaworabhan

2010

2010 Data Compression Conference

Self Cite

View full text Add to dashboard Cite

show abstract

“…Most of these parallelized methods divide a whole input stream into several fragments and perform the original sequential algorithm for each fragment in parallel. However, because predictors used in such techniques, e.g., FCM and DFCM, depend more or less on all previous inputs, dividing the input stream some- times diminishes prediction efficiency as mentioned in [10]. That is, we have a trade-off problem in parallel FP data compression between performance improvements by parallel effects and degradation of compression ratio resulting from lowered prediction efficiency.…”

Section: Preliminariesmentioning

confidence: 98%

“…A few techniques parallelizing these methods have been proposed [10], [18]. Most of these parallelized methods divide a whole input stream into several fragments and perform the original sequential algorithm for each fragment in parallel.…”

Section: Preliminariesmentioning

confidence: 99%

“…Most of all, these 1-pass algorithms are designed for serial processing, thus they are not easy to implement them on massively-parallel comput- [10], it is suitable for computers having some multi-core processors rather than massively-parallel systems. Therefore, recently O'Neil and Burtscher proposed another parallel FPC named GFC whose prediction algorithm is thoroughly redesigned for GPGPU [11].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Lossless Compression of Double-Precision Floating-Point Data for Numerical Simulations: Highly Parallelizable Algorithms for GPU Computing

Ohara

Yamaguchi

2012

IEICE Trans. Inf. & Syst.

View full text Add to dashboard Cite

SUMMARYIn numerical simulations using massively parallel computers like GPGPU (General-Purpose computing on Graphics Processing Units), we often need to transfer computational results from external devices such as GPUs to the main memory or secondary storage of the host machine. Since size of the computation results is sometimes unacceptably large to hold them, it is desired that the data is compressed and stored. In addition, considering overheads for transferring data between the devices and host memories, it is preferable that the data is compressed in a part of parallel computation performed on the devices. Traditional compression methods for floating-point numbers do not always show good parallelism. In this paper, we propose a new compression method for massivelyparallel simulations running on GPUs, in which we combine a few successive floating-point numbers and interleave them to improve compression efficiency. We also present numerical examples of compression ratio and throughput obtained from experimental implementations of the proposed method runnig on CPUs and GPUs.

show abstract

“…This increases the likelihood that, for an n-dimensional data set, every n th value be more likely to yield a reasonable prediction for the residual calculation and leading-zero compression. For instance, the CPU-based pFPC floating-point data compressor achieves better compression ratios when the number of threads is a multiple of the dimensionality of the data set [5]. Hence, GFC takes an optional dimensionality parameter between 1 and 32 to improve its performance.…”

Section: The Gfc Algorithm 21 Algorithmmentioning

confidence: 99%

Floating-point data compression at 75 Gb/s on a GPU

O'Neil

Burtscher

2011

Proceedings of the Fourth Workshop on General Purpose Processing on Graphics Processing Units

Self Cite

View full text Add to dashboard Cite

Numeric simulations often generate large amounts of data that need to be stored or sent to other compute nodes. This paper investigates whether GPUs are powerful enough to make real-time data compression and decompression possible in such environments, that is, whether they can operate at the 32-or 40-Gb/s throughput of emerging network cards. The fastest parallel CPUbased floating-point data compression algorithm operates below 20 Gb/s on eight Xeon cores, which is significantly slower than the network speed and thus insufficient for compression to be practical in high-end networks. As a remedy, we have created the highly parallel GFC compression algorithm for double-precision floating-point data. This algorithm is specifically designed for GPUs. It compresses at a minimum of 75 Gb/s, decompresses at 90 Gb/s and above, and can therefore improve internode communication throughput on current and upcoming networks by fully saturating the interconnection links with compressed data.

show abstract

pFPC: A Parallel Compressor for Floating-Point Data

Cited by 13 publications

References 14 publications

gFPC: A Self-Tuning Compression Algorithm

gFPC: A Self-Tuning Compression Algorithm

Lossless Compression of Double-Precision Floating-Point Data for Numerical Simulations: Highly Parallelizable Algorithms for GPU Computing

Floating-point data compression at 75 Gb/s on a GPU

Contact Info

Product

Resources

About