Use cases of lossy compression for floating-point data in scientific data sets

Cappello, Franck; Di, Sheng; Li, Sihuan; Liang, Xin; Gok, Ali Murat; Tao, Dingwen; Yoon, Chun Hong; Wu, Xin Chuan; Alexeev, Yuri; Chong, Frederic T.

doi:10.1177/1094342019853336

Cited by 118 publications

(59 citation statements)

References 56 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Unbiased compression methods which stay within the data's noise or analysis's error margin can be applied without degradating quality [40,73]. More use cases for lossy compression include the reduction of I/O time or the acceleration of checkpoint handling [30]. Lossy compression methods are usually used in a multilayer-compression approach, though specific algorithms reduce the data size sufficiently on their own: First, lossy compression reduces the data diversity so that the lossless compressors applied afterwards work more efficiently.…”

Section: Lossy Compressionmentioning

confidence: 99%

See 1 more Smart Citation

State of the Art and Future Trends in Data Reduction for High-Performance Computing

2020

JSFI

View full text Add to dashboard Cite

Research into data reduction techniques has gained popularity in recent years as storage capacity and performance become a growing concern. This survey paper provides an overview of leveraging points found in high-performance computing (HPC) systems and suitable mechanisms to reduce data volumes. We present the underlying theories and their application throughout the HPC stack and also discuss related hardware acceleration and reduction approaches. After introducing relevant use-cases, an overview of modern lossless and lossy compression algorithms and their respective usage at the application and file system layer is given. In anticipation of their increasing relevance for adaptive and in situ approaches, dimensionality reduction techniques are summarized with a focus on non-linear feature extraction. Adaptive approaches and in situ compression algorithms and frameworks follow. The key stages and new opportunities to deduplication are covered next. An unconventional but promising method is recomputation, which is proposed at last. We conclude the survey with an outlook on future developments.

show abstract

Section: Lossy Compressionmentioning

confidence: 99%

“…However, their compression scheme suffers from lower (de-)compression speed and disadvantageous random access times in return. Another drawback is the insufficient support for 2D datasets [30].…”

Section: Selection Of Lossy Compressorsmentioning

confidence: 99%

State of the Art and Future Trends in Data Reduction for High-Performance Computing

2020

JSFI

View full text Add to dashboard Cite

show abstract

“…If compression data can be kept completely in memory, out-ofcore algorithms can even be turned to in-core algorithms. A recent survey of use cases for reducing or avoiding the I/O bandwidth and capacity requirements in high performance computing, including results using mostly SZ and zfp, is given by Cappello et al [48].…”

Section: Compression Speed and Complexity Follow The Memory Hierarchymentioning

confidence: 99%

“…Consequently, only a moderate, but nevertheless consistent, benefit of compression has been shown in the literature. The broad spectrum of partially contradicting requirements faced by compression schemes in PDE solvers suggests that no single compression approach will be able to cover the need, and that specialized and focused methods will increasingly be developed-a conclusion also drawn in [48].…”

Section: Computational Fluid Dynamicsmentioning

confidence: 99%

Compression Challenges in Large Scale Partial Differential Equation Solvers

Götschel

Weiser

2019

Algorithms

View full text Add to dashboard Cite

Solvers for partial differential equations (PDE) are one of the cornerstones of computational science. For large problems, they involve huge amounts of data that needs to be stored and transmitted on all levels of the memory hierarchy. Often, bandwidth is the limiting factor due to relatively small arithmetic intensity, and increasingly so due to the growing disparity between computing power and bandwidth. Consequently, data compression techniques have been investigated and tailored towards the specific requirements of PDE solvers during the last decades. This paper surveys data compression challenges and discusses examples of corresponding solution approaches for PDE problems, covering all levels of the memory hierarchy from mass storage up to main memory. Exemplarily, we illustrate concepts at particular methods, and give references to alternatives.

show abstract

“…This throughput is far from enough for extreme-scale applications or advanced instruments with extremely high data acquisition rates, which is a major concern for corresponding users. The LCLS-II laser [10], for instance, may produce data at a rate of 250 GB/s [11], such that corresponding researchers require an extremely fast compression solution that can still have relatively high compression ratios-for example, 10:1-with preserved data accuracy. In order to match such a high data production rate, leveraging multiple graphics processing units (GPUs) is a fairly attractive solution because of its massive single-instruction multiple-thread (SIMT) mechanism and its high programmability as opposed to FPGAs or ASICs [12].…”

Section: Introductionmentioning

confidence: 99%

cuSZ

Tian

Zhao

et al. 2020

Proceedings of the ACM International Conference on Parallel Architectures and Compilation Techniques

Self Cite

View full text Add to dashboard Cite

Error-bounded lossy compression is a state-of-the-art data reduction technique for HPC applications because it not only significantly reduces storage overhead but also can retain high fidelity for postanalysis. Because supercomputers and HPC applications are becoming heterogeneous using accelerator-based architectures, in particular GPUs, several development teams have recently released GPU versions of their lossy compressors. However, existing state-of-the-art GPU-based lossy compressors suffer from either low compression and decompression throughput or low compression quality. In this paper, we present an optimized GPU version, cuSZ, for one of the best error-bounded lossy compressors-SZ. To the best of our knowledge, cuSZ is the first error-bounded lossy compressor on GPUs for scientific data. Our contributions are fourfold. (1) We propose a dual-qantization scheme to entirely remove the data dependency in the prediction step of SZ such that this step can be performed very efficiently on GPUs. (2) We develop an efficient customized Huffman coding for the SZ compressor on GPUs. (3) We implement cuSZ using CUDA and optimize its performance by improving the utilization of GPU memory bandwidth. (4) We evaluate our cuSZ on five real-world HPC application datasets from the Scientific Data Reduction Benchmarks and compare it with other state-of-the-art methods on both CPUs and GPUs. Experiments show that our cuSZ improves SZ's compression throughput by up to 370.1× and 13.1×, respectively, over the production version running on single and multiple CPU cores, respectively, while getting the same quality of reconstructed data. It also improves the compression ratio by up to 3.48× on the tested data compared with another state-of-the-art GPU supported lossy compressor.

show abstract

Use cases of lossy compression for floating-point data in scientific data sets

Cited by 118 publications

References 56 publications

State of the Art and Future Trends in Data Reduction for High-Performance Computing

State of the Art and Future Trends in Data Reduction for High-Performance Computing

Compression Challenges in Large Scale Partial Differential Equation Solvers

cuSZ

Contact Info

Product

Resources

About