On Mar 4, 10:21pm, Peter Teichmann wrote:
> Subject: Re: Once again: 16bpp setting for X
> > On Tue, 4 Mar 1997, Ale Terlevich wrote:
> >
> > Doesn't a 2MB VRAM machine have twice the max Video memory bandwidth to
> > screen of a 1MB VRAM machine, so you wouldn't expect the same mode to
> > necessarily work just because it fits into 1MB of RAM.
>
> Yes, that is certainly true. But I only have an AKF50, and so I have no
> problems with too high memory bandwidths needed. I only tried to use
> 800x600 pixels in 32000 colours. With editing the Bandlimit file in
> RiscOS I can achive 64 frames per second. But in RiscBSD it simply shows
> the rubbish I described in the last mail. I also reverted to 56 frames
> and then to 640x480 pixels which is definately no problem.
>
> Also I know how it looks if the VRam can not achive the bandwith I need
> from my experiments under RiscOS. But, as you can see from my last mail,
> it looks very different. Perhaps in 32000 colours the VIDC always becomes
> programmed  as if there were 2MB of VRam. I will test this as soon as I
> get some 2MB VRam to test. To buy them is to expensive for me and brings too
> less.

I don't know what is wrong with the old console code (it is odd because the
author had 1Meg VRAM). In theory the new console code should have no problems
with 1Meg VRAM but I have no easy way of testing it. If I bung you a test
kernel (it doesn't boot properly but text mode can run in 16bpp and 1bpp in
addition to 8bpp) could you test it for me?

Cheers

Rob