Georg Schönberger <gschoenberger@xxxxxxxxxxxxxxxx> writes: > Hi Jens, > > in the fio man-page you are writing: > "Even async engines my impose OS restrictions causing the desired > depth not to be achieved.This may happen on Linux when using libaio > and not setting direct=1, since buffered IO is not async on that OS." > (there's also a typo, I think it should be "may" instead of "my") > > Nevertheless it is possible to use fio with libaio and buffered IO: > $ fio --rw=write --name=test --size=20M --runtime=60 --time_based --iodepth=64 --ioengine=libaio > [...] > IO depths : 1=0.1%, 2=0.1%, 4=0.1%, 8=0.2%, 16=0.3%, 32=0.6%, >=64=98.8% > > Could you point out the limitations that come up with > libaio<->buffered IO (page cache)? The I/O is synchronous. In other words, if you issue io_submit with 32 iocbs, io_submit will not return until all 32 I/Os are complete (or there was an error). Cheers, Jeff -- To unsubscribe from this list: send the line "unsubscribe fio" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
