linux - Benchmarking PHP through unix socket vs tcp with ab and wrk -

when benchmarking regular php 5.6 through unix socket results many order of magnitues better tcp port.

when run command this:

$ ab -k -n 10000 -c 1000 http://127.0.0.1/api/user/1

i avg 3272 reqs per second.

but tcp port instead of unix socket 6.5 reqs per second.

with wrk

$ wrk -t1 -c1000 -d5s http://127.0.0.1:80/api/user/1

on unix socket: 6500 req per second

on tcp port: 300 req per second

how supposed use these benchmarks feel of how server , code can handle load when these kinds of results?

should trust tcp port or unix socket one?

how using unix socket in example? -k option http keep alive.

unix sockets used inter-process communication, i'm primary access method used access webserver via tcp.

really testing efficiency of code , webserver. if care efficiency of code, should consider inspecting xdebug output also.

JVParth