Search for: All records

Error-bounded lossy compression has been a critical technique to significantly reduce the sheer amounts of simulation datasets for high-performance computing (HPC) scientific applications while effectively controlling the data distortion based on user-specified error bound. In many real-world use cases, users must perform computational operations on the compressed data. However, none of the existing error-bounded lossy compressors support operations, inevitably resulting in undesired decompression costs. In this paper, we propose a novel error-bounded lossy compressor (called SZOps), which supports not only error-bounding features but efficient computations (including negation, scalar addition, scalar multiplication, mean, variance, etc.) on the compressed data without the complete decompression step, which is the first attempt to the best of our knowledge. We develop several optimization strategies to maximize the overall compression ratio and execution performance. We evaluate SZOps compared to other state-of-the-art lossy compressors based on multiple real-world scientific application datasets. 

Scientific simulations running on HPC facilities generate massive amount of data, putting significant pressure onto supercomputers’ storage capacity and network bandwidth. To alleviate this problem, there has been a rich body of work on reducing data volumes via error-controlled lossy compression. However, fixed-ratio compression is not very well-supported, not allowing users to appropriately allocate memory/storage space or know the data transfer time over the network in advance. To address this problem, recent ratio-controlled frameworks, such as FXRZ, have incorporated methods to predict required error bound settings to reach a user-specified compression ratio. However, these approaches fail to achieve fixed-ratio compression in an accurate, efficient and scalable fashion on diverse datasets and compression algorithms. This work proposes an efficient, scalable, ratio-controlled lossy compression framework (CAROL). At the core of CAROL are four optimization strategies that allow for improving the prediction accuracy and runtime efficiency over state-of-the-art solutions. First, CAROL uses surrogate-based compression ratio estimation to generate training data. Second, it includes a novel calibration method to improve prediction accuracy across a variety of compressors. Third, it leverages Bayesian optimization to allow for efficient training and incremental model refinement. Forth, it uses GPU acceleration to speed up prediction. We evaluate CAROL on four compression algorithms and six scientific datasets. On average, when compared to the state-of-the-art FXRZ framework, CAROL achieves 4 × speedup in setup time and 36 × speedup in inference time, while maintaining less than 1% difference in estimation accuracy.