[ardour-dev] Ancient history? - Problem identified

[Mark Knecht]

On Thu, 11 Nov 2004 13:31:32 -0800, David Gatwood <dgatwood at mac.com> wrote:
> 
> 
> 
> On Nov 11, 2004, at 12:40 PM, Jan Depner wrote:
> 
> > On Thu, 2004-11-11 at 11:56, Mark Knecht wrote:
> >
> >>    So, my first, somewhat obvious thought, is that drawing pink boxes
> >> is a process of moving a number of pixels to the screen. I presume you
> >> are using a more modern AGP based archtecture? If you are running an
> >> older PCI graphics adapter then you're moving those pixels across the
> >> same bus that your audio is moving across to go to and from the disk
> >> which would probably cause some problems.
> >>
> >       Even using AGP you're running over the PCI bus.
> 
> Well, more accurately, AGP is running over -a- PCI bus.  Not the same
> one as your sound card or SCSI card, though.

Sort of. Agp devices run PCI configuration cycles. However AGP is a
'split transaction' bus which allow the processor (be it the host CPU
or the graphics processor) to decouple the request from the action.
this allows a system to increase performance by executing commands out
of order and multiple commands at the same time.
> 
> AGP has to be on its own PCI-based bus because the clock rate is
> different than stock PCI....  Well, AGP 1x is 66 Mhz, so I guess in
> theory you could run AGP 1x over a slightly modified 66 MHz PCI bus,
> but AGP hardware can't necessarily fall back to 33 MHz speeds, so it
> would be a train wreck if you designed the hardware to share a single
> bus with most PCI devices....
> 
> However, you're right that AGP is basically just PCI at a potentially
> different clock speed multiplier with some extra bus bursting stuff
> hacked on top.  Kind of cool, as it means that you can write a single
> driver for AGP that works correctly for PCI devices as long as you
> don't check the error return from enabling the AGP-specific bus modes.
> :-D
> 
> But I digress.
> 
> 
> David
> 
>