Skip to content

Instantly share code, notes, and snippets.

@knutin
Last active April 8, 2017 16:59
Show Gist options
  • Select an option

  • Save knutin/78aca22b81cdfb537f4545913912e087 to your computer and use it in GitHub Desktop.

Select an option

Save knutin/78aca22b81cdfb537f4545913912e087 to your computer and use it in GitHub Desktop.
i3.2xlarge, 250M keys, values randomly distributed between 1024 and 4096 bytes. 596GB on NVMe drive.
Two r3.4xlarge clients are able to fully saturate the CPU on the server. iotop shows ~500MB/s read from the drive, iftop shows the same sent over the network.
With a thread for each physical core (i3.2xlarge has 4 cores and 8 processing units in total):
============================
Concurrent clients: 100
Runtime: 120 s
Total requests: 9468513
Avg rps: 78904.27
Bytes transferred: 23112.35 MB
Bytes per second: 192.60 MB
Roundtrip latencies: 50th: 1233us 75th: 1467us 90th: 1682us 95th: 1820us 99th: 2397us 99.9: 3534us
============================
Concurrent clients: 100
Runtime: 120 s
Total requests: 9502492
Avg rps: 79187.43
Bytes transferred: 23192.37 MB
Bytes per second: 193.27 MB
Roundtrip latencies: 50th: 1209us 75th: 1448us 90th: 1666us 95th: 1803us 99th: 2372us 99.9: 3424us
With a thread for each processing unit:
Client #1:
============================
Concurrent clients: 100
Runtime: 120 s
Total requests: 11471015
Avg rps: 95591.79
Bytes transferred: 28001.09 MB
Bytes per second: 233.34 MB
Roundtrip latencies: 50th: 1027us 75th: 1141us 90th: 1257us 95th: 1334us 99th: 1514us 99.9: 2952us
Client #2:
============================
Concurrent clients: 100
Runtime: 120 s
Total requests: 11471208
Avg rps: 95593.40
Bytes transferred: 28001.95 MB
Bytes per second: 233.35 MB
Roundtrip latencies: 50th: 1025us 75th: 1139us 90th: 1253us 95th: 1331us 99th: 1532us 99.9: 3110us
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment