Session: MCHPC’20: Workshop on Memory Centric High-Performance Computing

Architecting Heterogeneous Memory Systems with DRAM Technology Only: A Case Study on Relational Database

Wednesday, November 11, 2020

1:00 PM – 1:30 PM ET

Channel: Track 5

Author/Presenter(s)

YQ

Yifan Qiao

Rensselaer Polytechnic Institute, United States of America
XC

Xubin Chen

Rensselaer Polytechnic Institute (RPI), United States of America
JH

Jingpeng Hao

Rensselaer Polytechnic Institute (RPI), United States of America
TZ

Tong Zhang

Rensselaer Polytechnic Institute (RPI), United States of America
CX

Changsheng Xie

Huazhong University of Science & Technology, China
FW

Fei Wu

Huazhong University of Science & Technology, China

This paper advocates a DRAM-only design strategy to reduce the computing system memory cost and demonstrates its potential in the context of relational database. In particular, we envision a low-cost heterogeneous DRAM fabric consisting of convenient but expensive byte-addressable DRAM and large-capacity low-cost DRAM with coarse access granularity (e.g., 1K-byte). Regardless of specific memory technology, one can reduce the manufacturing cost by sacrificing the memory raw reliability, and apply error correction code~(ECC) to restore the data storage integrity. The efficiency of ECC significantly improves as the codeword length increases, which enlarges the memory access granularity. This leads to a fundamental trade-off between memory cost and access granularity. Following this principle, Intel 3DXP-based Optane memory DIMM internally operates with a 256-byte ECC codeword length (hence 256-byte access granularity), and Hynix recently demonstrated low-cost DRAM DIMM with a 64-byte access granularity. This paper presents a design approach that enables relational database to take full advantage of the envisioned low-cost heterogeneous DRAM fabric to improve performance with only minimal database source code modification. Using MySQL as a test vehicle, we implemented a prototyping system, on which we have demonstrated its effectiveness under TPC-C and Sysbench OLTP benchmarks.