Skip to content
This repository was archived by the owner on Mar 7, 2023. It is now read-only.

Throughput comparison

Jerrin Shaji George edited this page Mar 20, 2019 · 1 revision

KV pairs were inserted into the database using memtier benchmark. The below table reports throughput observed while varying the number of threads and number of clients per thread in memtier benchmark. Data objects inserted into the DB were sized between 32 and 1000 bytes. The throughput captured is average throughput over a period of 2 minutes.

Configuration

C Redis

  • Persistence mode - AOF with data fsynced to disk every second
  • Data file stored in persistent memory used as disk (block I/O mode)

Go Redis

  • Data stored in persistent memory used in byte addressable mode
Num. Threads Num. Clients C Redis (MB/sec) Go Redis (MB/sec) Speedup
1 50 166.8 403.66 2.42
2 50 173.31 429.14 2.47
3 50 173.12 406.36 2.34
4 50 170.35 407.22 2.39
Clone this wiki locally