FZ-GPU: A Fast and High-Ratio Lossy Compressor for Scientific Computing Applications on GPUs
FZ-GPU: A Fast and High-Ratio Lossy Compressor for Scientific Computing Applications on GPUs
复制标题
DOI：
10.1145/3588195.3592994
复制
发表时间：
2023-04
期刊：
影响因子：
--
通讯作者：
Bo Zhang;Jiannan Tian;S. Di;Xiaodong Yu;Yunhe Feng;Xin Liang;Dingwen Tao;F. Cappello
中科院分区：
文献类型：
--
作者：
Bo Zhang;Jiannan Tian;S. Di;Xiaodong Yu;Yunhe Feng;Xin Liang;Dingwen Tao;F. Cappello
如今，在高性能计算(HPC)系统上运行的大型科学应用程序会产生海量数据。因此，数据压缩成为减轻存储负担和数据移动成本的关键技术。然而，现有的科学数据有损压缩器不能同时实现高压缩比和高吞吐量，这阻碍了它们在许多需要快速压缩的应用中的采用，例如内存压缩。为此，在本工作中，我们在GPU上开发了一种快速、高误差率有界的科学数据有损压缩算法(简称FZ-GPU)。具体地说，我们首先设计了一种新的压缩流水线，它由完全并行化的量化、比特混洗和我们新设计的快速编码组成。然后，我们针对流水线中的每个内核提出了一系列深层次的架构优化，以充分利用CUDA架构的优势。我们提出了一种WARP级的优化方法，通过融合不同的压缩内核来避免位操作中的数据冲突，最大化共享内存利用率，并消除不必要的数据移动。最后，我们在两个NVIDIA图形处理器(即A100和RTX A4000)上使用来自SDRBch的六个具有代表性的科学数据集对FZ-GPU进行了评估。在A100图形处理器上的实验结果表明，在相同的误差界下，FZ-GPU在CuSZ上的平均加速比为4.2倍，在多线程CPU上的平均加速比为37.0倍。在相同的数据失真下，FZ-GPU与cuZFP相比，平均加速比提高了2.3倍，平均压缩比提高了2.0倍。
Today's large-scale scientific applications running on high-performance computing (HPC) systems generate vast data volumes. Thus, data compression is becoming a critical technique to mitigate the storage burden and data-movement cost. However, existing lossy compressors for scientific data cannot achieve a high compression ratio and throughput simultaneously, hindering their adoption in many applications requiring fast compression, such as in-memory compression. To this end, in this work, we develop a fast and high- ratio error-bounded lossy compressor on GPUs for scientific data (called FZ-GPU). Specifically, we first design a new compression pipeline that consists of fully parallelized quantization, bitshuffle, and our newly designed fast encoding. Then, we propose a series of deep architectural optimizations for each kernel in the pipeline to take full advantage of CUDA architectures. We propose a warp-level optimization to avoid data conflicts for bit-wise operations in bitshuffle, maximize shared memory utilization, and eliminate unnecessary data movements by fusing different compression kernels. Finally, we evaluate FZ-GPU on two NVIDIA GPUs (i.e., A100 and RTX A4000) using six representative scientific datasets from SDRBench. Results on the A100 GPU show that FZ-GPU achieves an average speedup of 4.2× over cuSZ and an average speedup of 37.0× over a multi-threaded CPU implementation of our algorithm under the same error bound. FZ-GPU also achieves an average speedup of 2.3× and an average compression ratio improvement of 2.0× over cuZFP under the same data distortion.