Hi all,

After a short conversation with Timothy about the possible availability of a PCI64 version of this project he asked me to continue the discussion on the mailinglist.

This was my introduction message and reason for this question:
Recently your attempt to create an "Open Source" 3d accelerated video adapter came to my attention and it left me with one question. It is probably not generic enough to post to the list, but I am using a Silicon Graphics O2 and (while the OpenGL performance was great in it's early days) it's performance starts to get slow compared to todays standards.
>
It would be great to be able to use your hardware as a replacement (or even add-on) for this machine. One advantage may be it's support for 64-bit PCI. Are you planning (or is it even possible) to support 64-bit PCI using this chip?

What follows is the largest part of the discussion between Timothy and me:

Erik Hofman wrote:

Timothy Miller wrote:

I've designed a 64-bit PCI controller before, so I know what's involved. It's just one of those cases where I know the transistor budget is really tight, so I don't even want to mess with it.

What's the demand anyhow? If PCIX is also 64-bit, and there's demand for that, we could meet in the middle just support PCI64/66.

I think PCIX is 64bit, isn't it? That would be a nice trade off.


This isn't impossible or anything, just generally undesirable for most users. It seems somewhat unproductive to put such a high-end feature on what's intended to be a relatively low-end card. Besides, I don't think the PCI bus will be the major bottleneck for scenes above a certain geometric complexity... then, the CPU becomes the bottleneck.

It depends on the application, most simulators rely on (lots of) texture swapping and try to get the number of drawn polygons as low as possible using techniques like impostors. In those circumstances bandwidth is the key issue.

Erik _______________________________________________ Open-graphics mailing list [email protected] http://lists.duskglow.com/mailman/listinfo/open-graphics List service provided by Duskglow Consulting, LLC (www.duskglow.com)

Reply via email to