spion / hashtable-latencies

Latency of a web service that stores a large hashtable, in multiple languages
43 stars 8 forks source link

python uwsgi #10

Closed gdamjan closed 8 years ago

gdamjan commented 8 years ago

single process, single threaded, no http keep-alive

VmPeak: 53620 kB VmRSS: 10660 kB

Thread Stats Avg Stdev Max +/- Stdev Latency 0.89ms 469.39us 11.20ms 68.99% Req/Sec 3.12k 183.66 5.80k 75.15%

still strange benchmark, http has too much effect. http/1.1 keepalive was worse for me, probably since there's only one socket accept, and wrk2 expects concurency. wrk2 uses more cpu than the app too