On 30 Jan 2014, at 21:38, Amos Jeffries <squ...@treenet.co.nz> wrote:
> On 2014-01-31 04:24, w...@wiki.squid-cache.org wrote: > > <snip> >> + || RPS || 39715 || >> + || Bandwidth || 313 Mbit/sec sustained || >> + > > ... "600-bytes long document". Thats 1KB per transaction. > >> + This number was taken in a '''controlled test environment'''. It has >> nothing to do with the numbers someone would get in a production >> environment; it's just an estimate of how fast squid can be. >> + Squid was configured to do no logging, no access control, and >> apachebench was used to hammer squid asking 10M times for a static, >> cacheable, 600-bytes long document. Of the 4 cores, 3 were running a >> multi-worker squid, one was running ab over the loopback interface. >> + > > Something seems to have slowed dow. I was getting roughly 8x that with > similar test environment (but single-core utilized by Squid) on 3.1 > transferring ~4KB objects. Although strangely it maxed at 100-200 Mbit/sec > for all the increased object size and throughput. > > What concurrency level was AB set to? I have seen quite a bit of variance in > speed by concurrency level. 100. I can do some random tests at different levels of concurrency, if useful. K
