A promise is a promise :) there you go!

Some shadermark scores would be nice if possible.

 

@ 16 pixel pipeline NV40, if that's true.

 

Vegetto,could you please confirm the VS/PS 3.0 support,if they're exposed through the current driver?
here's some shader model 3.0 demos:
http://www.beyond3d.com/forum/viewtopic.php?t=8789

 

This might be dumb, if so please forgive me , but could the prior cards be somewhat bandwidth limited... 1.2 ghz ram on the N40 would have given it a large advantage there.

 

Many are running 5950's above 1Ghz memory
and it does not pickup the score that much.

 

The bottleneck in a botmatch revolves a little bit more around the CPU than the graphics card. >_>

 

A graphic card does not double CPU limited botmatch scores, sorry.....

 

at this point, is there any hint of a doubt that NV40 is actually a 16 pipeline GPU?

I'm asking seriously.

I remember when everyone was certain NV30 had 8 pipelines. then a long while later, it was discovered NV30 was only a 4x2.....

 

IT can still be a 8x2 card for all we know .

I can do 16xcoreclock = fillrate for a 16x1 card and tell you what 3dmark will report in fillrate tests. Not that hard .

 

I dunno about you guys but to me 2.5-3 * or even 5 * increases over many of the low low nv3x scores while commendable, will still in many cases leave an nv40 barely beating a 9800xt.
Yes/No? :?

Doesn't nv3x have nice big fillrate test scores relative to r3x0 too?

 

Well, the cpu differences alone should put 60fps+ within reach. Also do the increases in cpu speed reflect linearly or non-linearly with regard to performance? Some simple testing could solve this. If its a linear progression in performance, then it would seem quite impossible. On the other hand...

EDIT: the 60fps+ figure is based on a linear assumption.... it would of course be more/less if this is not the case.

 

Are you suggesting an exponential increase in performance, or something?

How would that be possible?

 

Hey its all rumours there are some things that look that this could be a fake. IE the blurring in the screenshot when this is multi-sample either NV are already cheating by just blurring the screen or it is a normal vid card screenshot with a blurr applied.

We won't know until we get someone with more credible proof but that doesn't mean we have to murder him right now. Don't like it don't post and laugh at everyone afterwards if they are wrong.

 

1. No more tests. Vegetto doesn't have the card--he was our friendly neighborhood liaison to someone with a card. The card has since been sent back to wherever it goes, since he could only keep it for a few days. Sorry.

2. Leave the guy alone. He's from Chile, English is not his native language, and many of you are coming off very harsh. I've talked to him extensively; he's a good guy. He's not pulling anything.

3. Stuff I know. NV40 Ultra is clocked at 475/600, GDDR3, revision A2, will probably be an A3 at some point for power consumption, 16x1, very significant pixel shader improvements (guessing that for register-intensive FP32 shaders, we're looking at a 4-6x performance improvement versus NV35 clock-for-clock...), RGMS, adaptive AF (don't know if there will still be non-adaptive AF, we'll see), will ship with Forceware 60, review boards will be out end of March.

I'm still kind of curious about A3 and retail availability--I would imagine that it is coming to improve yields and reduce heat output (and the possible double molex connectors), but given the April 13 date, I'm forced to wonder about when it's going to be available.

(PS--it's 16x1. Period. There is no doubt.)

 

The review he posted to used a 3200. If his is overclocked to 3400+ that is only 200mhz .

Do you really think 2 ghz would double the framerate in cpu limited game .

That is just a huge increase for me to believe .

 

Again, those botmatch scores are simply not possible...end of story. So yes he is spreading lies...

 

That would be +400Mhz. The 3200+ runs at 2Ghz at stock.
________
METEOR

 

So your telling me i should believe double the framerate in botmatch ? That botmatch isn't cpu limited ?

 

Actually he has never had it now? and he is saying custom demos rather then the standard bot-match.

 

Well this is what is reflected in recent benchmark results for both the new P4 as well as the A64 (in some cases).

Additionally I was thinking that with a better cpu the video card would be doing more work too. I really have no idea though.