On Fri, 2015-09-18 at 22:24 +0200, Dragan Milivojević wrote: > > The throughput you're seeing with a single LUN + iodepth=1 + 1 Gb/sec is > > expected behavior. > > > > In order to saturate a 1 Gb/sec link, more outstanding I/Os need to be > > kept in flight (iodepth > 1) or a larger payload needs to be utilized. > > I just did a quick calculation (which I should have done before > writing this email) > If a total latency of a system is 1ms theoretical bandwidth comes down > to 1000 * 64K > which is around the figure I'm getting in the benchmarks. Little math > goes a long way :) > I should have done in the beginning and not waste this much time. > This should be in a wiki :) > > Thansk and sorry for wasting your time. > No worries. Thanks again for reporting the MSFT regression bug! --nab -- To unsubscribe from this list: send the line "unsubscribe target-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
