clusterperf benchmarks

clusterperf.py is a program that configures a RAMCloud cluster (using cluster.py) and then runs various benchmarks on it, using the program ClusterPerf as a client running on some of the cluster machines. This page contains the output of that program at various points in time, so it provides a history of RAMCloud cluster performance. Unless otherwise indicated, infrc is used as transport.

Note: for consistency, be sure to compile with DEBUG=no

• November 24, 2018 ({basic,homa}+infud) (Optimized infud bandwidth)

• November 10, 2018 ({basic,homa}+infud) (New transport: homa)

• August 23, 2016 (basic+dpdk) (DPDK over 1G Ethernet)

• January 6, 2016 (basic+infud) (New transport: basic)

• January 6, 2016 (Added linearizability support)

• January 30, 2015 (added dispatchExec, indexing)

• September 29, 2014 (Cache prefetching, various throughput optimizations, rewrite of multi-op client side)

• August 8, 2014 (Uses TCP over 1GigE, may be useful for others outside Stanford to compare against)

• June 19, 2014 (Buffer rewrite, ObjectFinder session caching)

• March 28, 2014 (on TCP/IP with -v option)

• August 13, 2013 (Table Stats Tracking)

• August 12, 2013

• November 12, 2012 (faster reads & multi-reads)

• June 20, 2012 (multi-read performance)

• April 12, 2012 (just after conversion to variable-length keys)

• September 13, 2011 (old 64-bit fixed-length keys)