PYXIS
收藏资源简介:
PYXIS是由加利福尼亚大学洛杉矶分校创建的一个开源性能数据集,专注于稀疏数据上的定制加速器。该数据集收集了加速器设计和实际执行性能统计,目前包含73.8千个实例。PYXIS使用来自SuiteSparse的2,637个稀疏矩阵,通过FPGA和GPU平台进行稀疏矩阵与密集矩阵的乘法运算,以收集性能数据。数据集的创建旨在解决稀疏应用加速器设计中缺乏性能分析模型的问题,适用于加速器、架构、性能和算法等多个研究领域。
PYXIS is an open-source performance dataset developed by the University of California, Los Angeles, focusing on custom accelerators for sparse data. This dataset collects accelerator design and actual execution performance statistics, and currently contains 73.8 thousand instances. PYXIS uses 2,637 sparse matrices from SuiteSparse, and conducts sparse-dense matrix multiplication on FPGA and GPU platforms to collect performance data. The dataset was created to address the lack of performance analysis models in the design of sparse application accelerators, and is applicable to multiple research fields including accelerators, computer architectures, performance evaluation, and algorithms.




