[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: NLR-AGP question



Doug Simmons wrote:
> Given: 64MB AGP dual-port video card
> Q: is the RAM used by each port limited to 32MB, or could one display be
> consuming 48MB and the other up to 16MB, for example?

No.  There is still 1 GPU.
The framebuffer is unified.
I.e., the screen is *always* rendered in a *fixed* geometry.
That's why you have to define where the 2nd head is, so the drive can 
organize the geometry.

E.g., if the 2 heads are side-to-side, and each head is 1600x1200, the 
framebuffer is 3200x1200.
Vertical would be 1600x2400.

Now that's the GPU.
It allows most GPUs to do fully accelerated 3D over 2+ heads.
Because it's just one big framebuffer, which all its pixel, texture, 
etc... units work on.

If you're talking Xinerama, that's different.
That's where the software combines separate framebuffers into one.
This could be on one card, or multiple cards.

BTW, the concept of the unified framebuffer has been around a long time.
The main reason why it became popular again was because the GDI in 
NT3/4, unlike OS/2, could not support multiple heads.
It wasn't until NT5 that Microsoft finally added support.
--
Bryan J. Smith   mailto:b.j.smith@ieee.org
Currently Mobile

-
To unsubscribe, send email to majordomo@silug.org with
"unsubscribe silug-discuss" in the body.