On Thu, 23 Apr 1998, Magnus Kollberg wrote:
> > I've had a quick look at it, and it looks rather interesting. I wish
> > NVDI could use this feature for caches...
>
> And if the AES could use it as well it would be REALY nice. There is no reason
> why not fVDI could make use of memory on the gfx card for offscreen bitmaps if
> there is enough memory available. On my 4Mb card there should be plenty of room
> in lower resolutions.
I discussed this with Geir Øyvind last weekend, and we came up with a
(possible) way of fooling NVDI to use videoRAM for it's buffers:
1. Run a program that allocates all available fast-RAM.
2. Use Maddalt() to register some of the videoRAM as alternate RAM
(e.g. the upper 1 Mb on a 2 or 4 Mb card).
3. Run NVDI, run in ST-RAM, Malloc from alternate RAM.
4. Release fast-RAM.
This way NVDI should use videoRAM for it's font-cache, and since it
most likely use a blit-function to render cached characters, it should
be a lot faster than usual (and reduce the load on the bus). But since
it must run entirely in ST-RAM the overall performance-increase will
probably not be dramatic.
I suggest that somebody with more than 1Mb videoram (I'm still looking
for a card with 2 or 4Mb VRAM...) tries something like this, it
shouldn't be more than a few lines of code.
#define MX_TTRAM 1
void main(void)
{
long *temp = NULL;
temp = Mxalloc(-1, MX_TTRAM);
Maddalt(VM_START, 1024*1024);
Pexec(0, "NVDI.PRX", NULL, NULL);
Mfree(temp);
}
VM_START points to the start of the last megabyte of videram, and can
be calculated by getting the base-address and ram-size from the
structure (XCB.H) pointed to by the NOVA-cookie.
Geir, do you have 10 minutes to spare tonight?
> BTW. Does the AES use the VDI to set up its screen buffer, or does it do it all
> by it self?
I don't know, but having the menu-buffer in videoRAM would be very
cool...
/*
** Jo Even Skarstein
http://www.stud.ntnu.no/~josk/
**
** beer - maria mckee - atari falcon - babylon 5
*/
Received on to. april 23 1998 - 16:31:16 CEST