[PDF][PDF] BigDataBench: a dwarf-based big data and AI benchmark suite

W Gao, J Zhan, L Wang, C Luo, D Zheng… - arXiv preprint arXiv …, 2018 - benchcouncil.org
… We need to understand What’s the abstractions of frequently-appearing units of computation
among big data and AI workloads (big data and AI dwarf)? … AI: ∎ ILP slightly lower than
SPECCPU, MLP similar with HPCC Big data has lower ILP ∎ and MLP than AI for almost all types,
except Hive based data warehouse type … Mainly ∎ DRAM bound for big data and AI benchmarks
∎ More stalls due to L1 Bound, L3 Bound and Store Bound than traditional benchmarks
以上显示的是最相近的搜索结果。 查看全部搜索结果