Boost logo

Boost :

From: Caleb Epstein (caleb.epstein_at_[hidden])
Date: 2005-12-28 12:04:02


On 12/28/05, Caleb Epstein <caleb.epstein_at_[hidden]> wrote:
>
>
>
> I'll see if I can't knock up a straight socket-based test like this one (
> e.g. just POSIX socket calls, no asio) and see if it gets similar results
> to Rene's benchmark. The same benchmark run on some other UNIX-based system
> (MacOS, *BSD, etc) would be another interesting data-point to have to see if
> this is an implementation or a platform issue.
>

OK, as threatened I've attached a POSIX-only version of Rene's test that
uses only socket, bind, select, sendto, recvfrom and friends. It looks
like, at least in the case of the asio ASYNC test, there is some performance
upside to be achieved:

Linux 2.4.21 SMP:

--- ASYNC...
### TIME: total = 1.30585; iterations = 100000; iteration = 1.30585e-05;
iterations/second = 76578.8
### TIME: total = 1.24593; iterations = 100000; iteration = 1.24593e-05;
iterations/second = 80261
### TIME: total = 1.24527; iterations = 100000; iteration = 1.24527e-05;
iterations/second = 80304.2
### TIME: total = 1.24671; iterations = 100000; iteration = 1.24671e-05;
iterations/second = 80211.2
### TIME: total = 0.995225; iterations = 100000; iteration = 9.95225e-06;
iterations/second = 100480
-- ...ASYNC: average iterations/second = 80258.8
--- SYNC...
### TIME: total = 1.10908; iterations = 100000; iteration = 1.10908e-05;
iterations/second = 90164.5
### TIME: total = 1.10027; iterations = 100000; iteration = 1.10027e-05;
iterations/second = 90886.4
### TIME: total = 1.1056; iterations = 100000; iteration = 1.1056e-05;
iterations/second = 90448.5
### TIME: total = 1.1051; iterations = 100000; iteration = 1.1051e-05;
iterations/second = 90490
### TIME: total = 1.10743; iterations = 100000; iteration = 1.10743e-05;
iterations/second = 90299.2
### TIME: total = 1.09628; iterations = 100000; iteration = 1.09628e-05;
iterations/second = 91217.8
-- ...SYNC: average iterations/second = 90531

Linux 2.6.13.4 uniprocessor:

--- ASYNC...
### TIME: total = 1.81951; iterations = 100000; iteration = 1.81951e-05;
iterations/second = 54960
### TIME: total = 1.78175; iterations = 100000; iteration = 1.78175e-05;
iterations/second = 56124.6
### TIME: total = 1.80059; iterations = 100000; iteration = 1.80059e-05;
iterations/second = 55537.3
### TIME: total = 1.81167; iterations = 100000; iteration = 1.81167e-05;
iterations/second = 55197.6
### TIME: total = 1.80915; iterations = 100000; iteration = 1.80915e-05;
iterations/second = 55274.6
### TIME: total = 2.77475; iterations = 100000; iteration = 2.77475e-05;
iterations/second = 36039.2
-- ...ASYNC: average iterations/second = 55533.5
--- SYNC...
### TIME: total = 1.48353; iterations = 100000; iteration = 1.48353e-05;
iterations/second = 67406.7
### TIME: total = 1.37942; iterations = 100000; iteration = 1.37942e-05;
iterations/second = 72494
### TIME: total = 1.36727; iterations = 100000; iteration = 1.36727e-05;
iterations/second = 73138.3
### TIME: total = 1.37749; iterations = 100000; iteration = 1.37749e-05;
iterations/second = 72595.8
### TIME: total = 1.35504; iterations = 100000; iteration = 1.35504e-05;
iterations/second = 73798.8
### TIME: total = 1.19966; iterations = 100000; iteration = 1.19966e-05;
iterations/second = 83357.3
-- ...SYNC: average iterations/second = 73006.7

--
Caleb Epstein
caleb dot epstein at gmail dot com



Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk