First NV31 Announcement Rumors...

[RussSchultz]

However it is also the analog parts that are most likely to get passed along between individual products, as long as they stick with a specific fab/process. All of the 0.15 Nvidia (or ATI) chips generally use the same process. When that is the case, the analog stuff is designed once for that process, and then left mostly as-is, unless they are broken or deficient in some way. So ATI's R200/RV250/R300 likely use many of the same analog parts (PLLs, DLLs, DACs, I/Os, etc...) because there is no reason to waste resources redesigning them. So if NV34 were to be a 0.15 version of some cut down NV30/31 logic, it's likely that many of the analog parts are pulled from the nv25/28. And future 0.13 designs from Nivida and ATI will probably share many of the analog parts with the nv30 and rv350.

I agree. But its not the analog parts that separate these chips' functionality between DX8 or DX9.

I'll defer to your estimates on synthesis time. (though I didn't suggest it was started at 80%, only that it took about 20% of the project's scheduled time.).

 

[Arun]

Never posted here before, mostly posted at nV News in the past. Thanks Dave Baumann for activating my account
Anyway, I've already posted this at nV News, but I think it makes more sense in this thread...

After a some calculations based on the 400/450 ( effective 400/900 ) GFFX Regular number, I estimated that the NV3x architecture needs only 56% of the memory frequency the NV2x architecture requires, if using DDR2 & all the memory saving techniques.

Thus, since Mufu is saying that the NV31 uses DDR1 & got some features removed, I'm guestimating it needs 75% of the memory frequency of the NV2x architecture. And what does that mean? That if we get a 400Mhz clock frequency for a 4x1 NV31, we'd get a 300Mhz memory frequency.

And that's exactly the same as the AGP 4X Ti4600. And at the same time, The Inquirer says the NV31 is a revamped Ti4600. Now, The Inquirer isn't a reliable source, but with the NV30, they got the memory speed right. Maybe that means they've got reliable sources in the memory area. Or maybe they got lucky
So, could it be they've got the 300Mhz memory frequency, and thus found it logical to be compared to a Ti4600?

Of course, that number is only correct if nVidia is able to get 400Mhz. But considering that 300Mhz memory seems very logical, and that's the way to make it balanced, it sounds good to me.

Anyway, I'd love to get a few clarifications on Mufu's comments. The 400/350 numbers, did he get it from his source or is it his best guess?
And "some functional blocks removed and compensated in software" - could it mean the whole feature set of CineFX isn't there, and that some instructions are automatically transformed into multiple others by the driver before sending it to the GPU? Several such instructions exist in the feature set, so it's certainly possible.
That would obviously be better than simply retrieving the whole VS... But then again, it would cause one of those mess for programmers, not knowing at all how it would perform on a NV31 before testing it...


Uttar

 

[Doomtrooper]

Uttar, maybe you could enlighten us with your comment...

I didn't say "losing money". I said they barely made any money on it. So, they'd barely make any money on high-end to get a good reputation and be able to sell their low-end ( Radeon 9500/9000 ) to make money
nVidia, on the other hands, prefer to make more money on high-end.

I found that above comment very funny, maybe you could explain..especially the 9500 being 'low end' when it currently runs circles around anything else out there...where are you getting your ATI financials from.

 

[Dave Baumann]

Lets keep the comments to those made in this forum / thread.