Re: benchmarking i/o throughput
Andreas Schäfer <[email protected]>
| Newsgroups | gmane.linux.cluster.openmosix.general |
|---|---|
| Message-ID | <20060226044733.GD16375@wintermute> |
> > gentryx@wintermute ~ $ ./writeTest 1024 1000 >/tmp/foo > > status: 0 blksize: 131072 > > sleeping 1000 seconds > > dumping 1024 mega bytes > > write took 387639 milli seconds > > 2.70503 MB/s Update: I can get even better rates for files < 1024 MB. Those files don't need to be written in place as there's enough buffer mem on my system. This allows rates of 3.4 MB/s. When writing into a pipe even higher rates are feasible. Per page my block device performs 3 accesses to the userspace memory: two times a deputy_get_userX() for different regions (I assume to preload the cachelines), followed by one deputy_get_user() for the actual page. The pipe leaves out one of the deputy_get_userX() calls which leads to a higher througput: 4.5135 MB/s. -Andreas ------------------------------------------------------- This SF.Net email is sponsored by xPML, a groundbreaking scripting language that extends applications into web and mobile media. Attend the live webcast and join the prime developer group breaking into this new coding territory! http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&dat=121642