Near-Data Filters: Taking Another Brick from the Memory Wall
In this paper, we use the potential of the near-data parallel computing presented in the Hybrid Memory Cube (HMC) to process near-data query lters and mitigate the data move- ment through the memory hierarchy up to the x86 processor. In particular, we present a set of extensions to the HMC In- struction Set Architecture (ISA) to lter data in-memory. Our near-data lters support vector instructions and solve data and control dependencies internally in the memory: in- ternal register bank and branch-less evaluation of data lters transform control- ow dependencies into data- ow depen- dencies (i.e., predicated execution). We implemented the near-data lters in the select scan operator and we discuss preliminary results for projection and join. Our experiments running the select scan achieve performance improvements of up to 5.64 with an average reduction of 80% in en- ergy consumption when executing a micro-benchmark of the 1 GB TPC-H database.
|Conference||Conference: Conference: ADMS/VLDB, At Rio de Janeiro|
Tomé, D.G, Kepe, T.R, Alves, M.A.Z, & de Almeida, E.C. (2018). Near-Data Filters: Taking Another Brick from the Memory Wall.