From: Stephen Wille Padnos <spadnos@sover.net>
To: Timothy Miller <miller@techsource.com>
Cc: Jon Smirl <jonsmirl@gmail.com>,
Linux Kernel Mailing List <linux-kernel@vger.kernel.org>
Subject: Re: HARDWARE: Open-Source-Friendly Graphics Cards -- Viable?
Date: Thu, 21 Oct 2004 12:34:23 -0400 [thread overview]
Message-ID: <4177E50F.9030702@sover.net> (raw)
In-Reply-To: <4177DF15.8010007@techsource.com>
Timothy Miller wrote:
> You're right, and you're not. There are two reasons why modern 3D
> GPUs put the world mesh into the card framebuffer and do all the T&L
> in the GPU. One reason is that it's faster to do it in dedicated
> hardware. The other more pressing reason is that the host interface
> (PCI or AGP) puts a very low upper-bound on your triangle rendering rate.
I'm thinking more like microcode. The functional blocks on the chip
would be capable of being "rewired" by the OS, depending on the
applications being run. All of the functions would still operate out of
card-local memory.
> The number of parameters necessary to specify a single texture-mapped
> triangle are literally in the dozens. If you did only 32-bit fixed
> point (16.16) for coordinates, that's still a huge amount of
> information to move across the bus to instruct the chip to render a
> single triangle. And what if that triangle ends up amounting to a
> single pixel? Think about the waste!
>
> Instead, the un-transformed geometry is loaded into the card memory
> only once, and the GPU is instructed to render the scene based on the
> camera perspective and lighting information. Aside from the need to
> cache textures on-card, this is another reason for the need to have
> LOTS of graphics memory.
Yep - reminds me of the days with the old SGI Iris, which did OpenGL
processing directly on the card, with the display list local to the
GPU(s) (1280x1024@100 bit depth -24 bit color, 24-bit Z, another 48 bits
for double-buffering, and 2 bits of overlay + 2 more underlay - those
were the days :)
>> I guess I would look at this as an opportunity to make a "visual
>> coprocessor", that also has the hardware necessary to output to a
>> monitor (preferably multiple monitors).
>
> I don't think that's realistic. We could do that, but it would have
> terrible performance.
Sorry - I wasn't thinking in terms of an 8087 - more like a dedicated
processor for image related functions - like a DSP card. A display list
based coprocessor would certainly be better than anything that has to
repeatedly transfer commands over a slow bus.
- Steve
next prev parent reply other threads:[~2004-10-21 16:41 UTC|newest]
Thread overview: 165+ messages / expand[flat|nested] mbox.gz Atom feed top
2004-10-20 22:02 HARDWARE: Open-Source-Friendly Graphics Cards -- Viable? Timothy Miller
2004-10-20 22:17 ` Andre Eisenbach
2004-10-21 1:31 ` Jon Valvatne
2004-10-21 16:09 ` Timothy Miller
2004-10-24 19:47 ` Pavel Machek
2004-10-20 22:19 ` HARDWARE: Open-Source-Friendly Graphics Cards -- Viable? [u] Martin Schlemmer [c]
2004-10-24 8:24 ` Tonnerre
2004-10-24 14:26 ` Martin Schlemmer [c]
2004-10-20 22:26 ` HARDWARE: Open-Source-Friendly Graphics Cards -- Viable? David Lang
2004-10-21 14:46 ` Timothy Miller
2004-10-21 17:25 ` David Lang
2004-10-21 18:15 ` Timothy Miller
2004-10-21 18:32 ` Antonio Vargas
2004-10-22 9:53 ` Raphael Jacquot
2004-10-24 9:03 ` Tonnerre
2004-10-25 1:33 ` Stephen Wille Padnos
2004-10-25 1:48 ` Stephen Wille Padnos
2004-10-25 2:29 ` Gene Heskett
2004-10-22 10:16 ` Christian Leber
2004-10-22 17:31 ` Timothy Miller
2004-10-21 19:30 ` Kendall Bennett
2004-10-22 17:05 ` Tobias Diedrich
2004-10-22 17:12 ` Timothy Miller
2004-10-26 2:36 ` Dave Airlie
2004-10-26 3:55 ` Jon Smirl
2004-10-20 22:28 ` Jim Nelson
2004-10-21 14:51 ` Timothy Miller
2004-10-21 22:03 ` Jim Nelson
2004-10-20 22:29 ` Kasper Sandberg
2004-10-21 14:53 ` Timothy Miller
2004-10-21 15:06 ` Simon Braunschmidt
2004-10-21 18:00 ` Timothy Miller
2004-10-20 23:10 ` Alan Cox
2004-10-21 15:10 ` Timothy Miller
2004-10-21 15:25 ` Jon Smirl
2004-10-21 18:03 ` Timothy Miller
2004-10-21 15:32 ` Alan Cox
2004-10-21 19:30 ` Kendall Bennett
2004-10-22 17:15 ` Timothy Miller
2004-10-21 1:08 ` Jon Smirl
2004-10-21 1:11 ` Jon Smirl
2004-10-21 2:00 ` Stephen Wille Padnos
2004-10-21 16:08 ` Timothy Miller
2004-10-21 16:34 ` Stephen Wille Padnos [this message]
2004-10-21 23:38 ` Jan Knutar
2004-10-22 4:30 ` Jan Rychter
2004-10-22 17:00 ` Timothy Miller
2004-10-22 17:00 ` Chris Friesen
2004-10-22 18:47 ` Jeff Garzik
2004-10-22 19:22 ` Timothy Miller
2004-10-22 19:33 ` Jeff Garzik
2004-10-22 19:56 ` Timothy Miller
2004-10-22 20:43 ` Jeff Garzik
2004-10-22 20:27 ` Alan Cox
2004-10-23 17:20 ` Francois Romieu
2004-10-23 21:17 ` Alan Cox
2004-10-24 0:06 ` Francois Romieu
2004-10-22 20:51 ` Jeff Garzik
2004-10-22 19:32 ` Roland Dreier
2004-10-24 10:40 ` Helge Hafting
2004-10-25 15:39 ` Timothy Miller
2004-10-21 21:57 ` J.A. Magallon
2004-10-22 9:48 ` Raphael Jacquot
2004-10-21 20:23 ` "Fernando O. Korndörfer"
2004-10-22 9:02 ` Raphael Jacquot
2004-10-21 15:13 ` Timothy Miller
2004-10-21 15:36 ` Shaun Kruger
2004-10-21 18:05 ` Timothy Miller
2004-10-21 19:30 ` Kendall Bennett
2004-10-22 8:49 ` Adrian Cox
2004-10-22 20:10 ` Geert Uytterhoeven
2004-10-23 13:17 ` Adrian Cox
2004-10-22 20:10 ` Geert Uytterhoeven
2004-10-22 22:07 ` Timothy Miller
2004-10-24 10:45 ` Helge Hafting
2004-10-25 15:47 ` Timothy Miller
2004-10-28 9:07 ` Helge Hafting
2004-10-29 16:00 ` Timothy Miller
2004-10-21 1:48 ` HARDWARE:Graphics Cards or TOE? Nuno Silva
2004-10-26 20:50 ` Timothy Miller
2004-10-21 2:29 ` HARDWARE: Open-Source-Friendly Graphics Cards -- Viable? Kurt Wall
2004-10-21 16:10 ` Timothy Miller
2004-10-21 16:22 ` Pascal Patry
2004-10-21 12:20 ` Adrian Bunk
2004-10-21 13:14 ` Simon Braunschmidt
2004-10-21 17:34 ` Jurriaan
2004-10-21 16:26 ` Timothy Miller
2004-10-21 17:42 ` Alan Cox
2004-10-21 19:09 ` Timothy Miller
2004-10-21 17:53 ` Tobias Diedrich
2004-10-21 23:02 ` Florian Schmidt
2004-10-24 1:04 ` Lee Revell
2004-10-22 1:08 ` Rene Herman
2004-10-23 5:40 ` Kevin Puetz
2004-10-23 17:02 ` Rene Herman
2004-10-23 22:19 ` Lee Revell
2004-10-24 11:10 ` Rene Herman
2004-10-22 10:57 ` Helge Hafting
2004-10-22 19:47 ` Giuseppe Bilotta
2004-10-22 20:15 ` Giuseppe Bilotta
2004-10-25 15:29 ` Tonnerre
2004-10-25 15:53 ` Timothy Miller
2004-10-25 16:32 ` Giuliano Pochini
2004-10-28 9:37 ` Helge Hafting
2004-10-28 11:40 ` Geert Uytterhoeven
2004-10-28 12:21 ` David Greaves
2004-10-29 16:04 ` Timothy Miller
2004-10-22 22:27 ` Clemens Schwaighofer
2004-10-23 14:36 ` Markus Törnqvist
2004-10-24 8:18 ` Tonnerre
2004-10-25 11:54 ` Stuart Longland
2004-10-25 16:38 ` Lars Roland
2004-10-25 17:08 ` Timothy Miller
2004-10-26 21:02 ` Helge Hafting
2004-10-26 21:38 ` Timothy Miller
2004-10-25 22:52 ` Tonnerre
-- strict thread matches above, loose matches on Subject: below --
2004-10-20 23:48 Timothy Miller
2004-10-21 0:30 ` J.A. Magallon
2004-10-21 0:47 ` Timothy Miller
2004-10-22 20:09 ` Geert Uytterhoeven
2004-10-21 1:25 ` Zan Lynx
2004-10-21 15:52 ` Timothy Miller
2004-10-21 4:48 Albert Cahalan
2004-10-21 16:19 ` Timothy Miller
2004-10-21 15:54 John Ripley
2004-10-21 18:09 ` Timothy Miller
2004-10-21 21:32 ` Baruch Even
2004-10-25 23:30 ` Werner Almesberger
2004-10-21 17:08 Greg Buchholz
2004-10-22 2:18 ` Tim Connors
2004-10-21 17:44 John Ripley
2004-10-21 18:26 ` Timothy Miller
2004-10-21 21:36 ` Greg Buchholz
2004-10-21 22:40 ` Timothy Miller
2004-10-21 23:25 ` Jon Smirl
2004-10-21 23:40 ` Greg Buchholz
2004-10-22 16:48 ` Timothy Miller
2004-10-22 16:50 ` Chris Friesen
2004-10-22 17:41 ` Timothy Miller
2004-10-25 23:10 ` Tonnerre
2004-10-26 0:32 ` Werner Almesberger
2004-10-22 15:59 ` Troy Benjegerdes
2004-10-22 3:47 Roy Butler
2004-10-22 17:04 ` Timothy Miller
2004-10-24 18:17 ` Mail Lists
2004-10-25 12:17 ` Bernd Eckenfels
2004-10-22 10:31 John Ripley
2004-10-22 12:58 ` Moritz Muehlenhoff
2004-10-22 17:33 ` Timothy Miller
2004-10-22 17:15 Stephen Lewis
2004-10-23 4:45 ` Gene Heskett
2004-10-23 7:06 ` Stephen Lewis
2004-10-23 19:06 Bodo Eggert
2004-10-25 1:44 ` Stephen Wille Padnos
2004-10-25 8:23 ` Vojtech Pavlik
[not found] <6.1.2.0.1.20041026082223.0231edd8@mail.javagear.com>
2004-10-26 15:44 ` Timothy Miller
2004-10-26 16:35 ` Jesper Juhl
2004-10-26 16:57 ` Jeff Garzik
2004-10-26 21:14 ` Helge Hafting
2004-10-26 21:41 ` Timothy Miller
2004-11-17 14:35 Sid Boyce
2004-11-17 14:46 ` Chris Wedgwood
2004-11-23 13:47 ` Karel Kulhavy
2004-11-23 22:48 ` Timothy Miller
2004-11-24 1:22 ` Sid Boyce
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=4177E50F.9030702@sover.net \
--to=spadnos@sover.net \
--cc=jonsmirl@gmail.com \
--cc=linux-kernel@vger.kernel.org \
--cc=miller@techsource.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).