"DBus Embedded" - a clean break

[Thiago Macieira]

On Thursday, 20 de January de 2011 14:18:14 Rob Taylor wrote:
> On 20/01/11 13:47, Thiago Macieira wrote:
> > On Thursday, 20 de January de 2011 10:09:00 Rob Taylor wrote:
> >> P.S.  Context for those not involved in embedded systems: consider
> >> that
> >> I generally estimate send->receive time of 30ms for a single dbus
> >> message going via the bus on capable ARM systems. That means the time
> >> for 10 IPC calls is a human-visible time.
> > 
> > My benchmarks on the N900 produce 4 to 5 ms for the roundtrip.
> > 
> > That's not just me either:
> > http://permalink.gmane.org/gmane.comp.freedesktop.dbus/10252
> 
> Good stuff, I've obviously been being too pessimistic! Are your bench
> marks measuring different message sizes and processor loads?

Yes. It shows that the payload size, under a certain threshold, does not 
influence the delivery time. Sending more data in a single message just 
increases the data throughput.

Then it plateaus and adding more data makes it slower. The overhead of parsing 
more 2048-byte blocks from the socket and allocating memory takes over.

-- 
Thiago Macieira - thiago (AT) macieira.info - thiago (AT) kde.org
  Senior Product Manager - Nokia, Qt Development Frameworks
      PGP/GPG: 0x6EF45358; fingerprint:
      E067 918B B660 DBD1 105C  966C 33F5 F005 6EF4 5358
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 190 bytes
Desc: This is a digitally signed message part.
URL: <http://lists.freedesktop.org/archives/dbus/attachments/20110120/14f92ef5/attachment.pgp>