On Tue, Oct 11, 2011 at 10:23 PM, Steven McCoy <[email protected]> wrote:
> Took me a while as I couldn't see it in Google Preview: > 999 instances of ~999ns or whatever the time scale is and a final one at > ~20,000ns. This is over TCP? This is probably an artefact of the way you're measuring latency. 0MQ batches messages aggressively so you're probably seeing two batches, with a single message in the second batch. Naturally it'll arrive with an apparent spike in latency. -Pieter _______________________________________________ zeromq-dev mailing list [email protected] http://lists.zeromq.org/mailman/listinfo/zeromq-dev
