About performance of D-Bus
rob.taylor at codethink.co.uk
Thu Oct 25 04:16:51 PDT 2007
Don't get me wrong! I'm not suggesting his results are bad, just want to
make sure we're comparing like with like before I start suggesting fixes!
Luca De Cicco wrote:
> The same delay was measured by me in an application I was writing... I
> then gave up using D-BUS for massive data communication between
> applications and failed-over to plain sockets. Btw the delay was quite
> the same that Tian is measuring with some spikes of 16ms.
> On Thu, 25 Oct 2007 09:17:17 +0100
> Rob Taylor <rob.taylor at codethink.co.uk> wrote:
>> You're not comparing like with like - The 2nd case would be more like
>> the first if you print the gettimeofday in the the dbus_filter
>> function. Could you do this and see how that compares? Also make sure
>> the first case is using a select rather than a blocking read and that
>> there's no other processes running for good reproducibility.
>> Rob Taylor
Rob Taylor, Codethink Ltd. - http://codethink.co.uk
More information about the dbus