Glamo Speed Improvement(s)

[Vasco Névoa]

Mamma mia!!!  :D
I'm salivating already, in anticipation of the day that SHR will  
include this kernel in it's updates. :D

Citando Thomas White <taw at bitwiz.org.uk>:

> On Sun, 15 Nov 2009 15:35:33 +0100
> Nicola Mfb <nicola.mfb at gmail.com> wrote:
>
>> I'm waiting for that great news for months!!!!
>> Hope to find time during the next week to revamp my build host and do
>> some testing ;)
>
> Great stuff - let me know what happens.  There's an overhead associated
> with sending the process to sleep and waking it up again, but you
> should find that CPU usage goes down quite a lot.  The best results are
> obtained when the commands being sent to Glamo are for fills/blits of
> larger areas.  The breakeven point "sadly" got a lot larger when the 2D
> engine doubled in speed.  Perhaps some kind of hybrid system would work
> best, with low-latency busywaits used for short waits and the waitqueue
> system used for longer ones.  As a general benchmark, my test program
> (flat out fullscreen solid fills, waiting each frame) gets around 96
> frames per second with about 3% CPU usage.
>
> Tom
>
> --
> Thomas White <taw at bitwiz.org.uk>
>
> _______________________________________________
> devel mailing list
> devel at lists.openmoko.org
> https://lists.openmoko.org/mailman/listinfo/devel
>