NV40 supports 3Dc

Sage said:
oh it was like over a year ago that HE said that
this has been going on over a year.

dammit, that sucks. so there are NO women on this board? pffft that sucks.
I still find this whole thing f'in hilarious. There's no way I ever said that, ever. But I still find the image very funny of this happening in the real world. I mean, I'm 6'4" with thinning hair and I don't shave religiously. I just found the thought of somebody claiming I was female was just utterly hilarious in relation to the real world. Not only that, but my name, which I swear I didn't select from here, but happens to be a character from star trek:
http://www.angelfire.com/nh/AvalonScience/images/Chalnoth.jpg
From:
http://www.angelfire.com/nh/AvalonScience/XenologicalListings01.html (google search)
 
Far Cry 1.3 detects 3Dc the same way The Project demo does
From description of Far Cry console cmds:
"Toggles support for normal-maps compression. 3Dc for ATI and V8U8 for NVidia".
So, I think, if NV card is detected, Far Cry 1.3 will use V8U8 anyway.
 
digitalwanderer said:
Chalnoth said:
DaveBaumann said:
I suspect that 512MB cards will be in the high end minority for a fair while as well!
Memory sizes tend to filter down to low-end parts more quickly than anything else in the high-end space, since there's still this perception that the amount of memory a video card has is the primary determination of performance.
Oh man, wasn't there a 256MB 5200 or something silly like that? :oops: :LOL:

http://www.google.com.au/search?num=100&hl=en&safe=off&q=radeon+9000+256mb&meta=
Something as silly as a Radeon 9000 256mb digi? :LOL: :rolleyes:
 
I really dont see your point radar. I dont think digi was taking a jab at Nvidia. Just low end cards with 256 megs onboard.
 
And the other point being that with "HyperMemory" / "TurboCache" they'll be more low end board with miniscule quantities of on-board memory, with the major texture cahce being in system RAM. Compression, for performance purposes, will be very useful in these situations.
 
I never said he was. And I never defended the 256mb 5200 either. Just pointing out nVidia isn't the only IHV to offer low end 256mb products.

a 256mb 5200 and 256mb 9000 are equally ridiculous IMO.

But I'll bet anything you like that if the original post had mentioned a 256mb 9000 instead of a 5200 digi wouldn't have said a word.
 
dose ATi or nVIDIA even have some thing to do with those cards?
some IHV wanted to make their card more appealing by shoving alot of cheap ram into them. so what dose it have to do with either ATi or nVIDIA. you can shove 256mb into a voodoo2 based card can`t you?
 
DOGMA1138 said:
dose ATi or nVIDIA even have some thing to do with those cards?
some IHV wanted to make their card more appealing by shoving alot of cheap ram into them. so what dose it have to do with either ATi or nVIDIA. you can shove 256mb into a voodoo2 based card can`t you?
I'm pretty sure you can't. The memory controller needs to be able to address so much memory (basically needs enough address lines). ATI/Nvidia certainly could restrict this, but it would be a "artificial" limitation. It certainly makes sense for the chip to be able to address anything you might throw at it during it's lifetime...

mczak
 
Hyp-X said:
Xmas said:
It works just like 3Dc, so developers don't need to treat R420 and NV40 differently. And it still only takes half the space of uncompressed 3(4)-channel normal maps.

Any developer who doesn't know how to use U8V8 (16bit, supported on all DX9 hardware) for uncompressed format won't be qualified to use 3Dc anyway.
I believe this feature is done primary so nVidia cards run ATi techdemos (so their cards doesn't look bad for the users).

R420 and NV40 should still be treated differently if you want the texture to occupy the same amount of memory (video memory usage being the topmost concern).

This trick actually makes right feature detection harder - just like when we had to add special cases for GF4MX (it said it can do VS) - or GF2 cards (they said the can do non-pow2 textures)
Developers have been using uncompressed 32bit textures as normal maps for a while. And they can't use 2-channel formats only, as long as there is sub-PS2.0 hardware. I think NVidia wants to avoid a situation where a developer only uses 3Dc in addition to uncompressed normal maps, or where there is a "fallback" to DXT5, because on NVidia hardware V8U8 is usually faster than using only G and A components of DXT5 compression, despite the bandwidth/memory disadvantage.
 
radar1200gs said:
But I'll bet anything you like that if the original post had mentioned a 256mb 9000 instead of a 5200 digi wouldn't have said a word.
Considering I'm the one who mentioned the 256Mb 5200 in the first place, I have NO CLUE what in the heck you are going on about! :?
 
Back
Top