The LAST R600 Rumours & Speculation Thread

Status
Not open for further replies.
I don't like to be wrong....

R480
Pixel Pipelines 16
1:1 Pixel Shaders 16

R580
Pixel Pipelines 16
3:1 Pixel Shaders 48
 
Dave Baumann said:
Kinda depends on how you look at things really.

Hark back to Voodoo days and we have separate chips for the pixel pipeline and the texture pipeline - they were decoupled then (with a scalable number of texture units). We've gone through a whole mess of stuff since then, but if you look at something like Xenos there are remarkable paralells in that the texture pipes and the pixel pipes are separate chips!
The "pixel pipeline" can just be boiled down to the ROP's, which are still relatively fixed function elements that they were in Voodoo days - the texture piplines, though, have grown to be rather large, programmble units.
Click to expand...

I would like to say thank you for your support my friend! Dave Baumann

rwolf said:
Last I checked R580 had a 16 pixel pipeline with 48 pixel processors. Just because the implementation is not always 1:1 for all components doesn't mean the pixel pipeline is not there.
Click to expand...

Also thank you for your comment: Rwolf

--------------------------------------------------------------------------------------------------

For all negative responds I had from people such as Uttar, Morgoth the Dark Enemy, etc.... Don't judge the book by its cover, I'm not totally stupid.... "I never said I was professional like other people, but I'm not stupid either"
 
Last edited by a moderator:
Jawed said:
I think Uttar is saying that each cluster, which has two 16KB PDCs, adds up to a total PDC size of 256KB.
Click to expand...
Yeah, that's what he seems to be saying. No matter how many PDCs there are though it seems that according to Nvidia 16KB is available to share between threads. The number of PDCs is only important for latency hiding as they and the software will dictate how many threads/batches can be in flight.
 
Unknown Soldier said:
Maybe that's the R630? I can't think the final R600 is only released now.
Click to expand...

It's approximately 6 weeks from now to the second week of March. How much more do they need for mass production? (granted there are also the 2 weeks halt for the chinese new year, but I don't think it'll change anything dramatically in the end).
 
Well, we've usually seen 3 months between tape-out and availability. I wouldn't know how much time is subtracted if they already have chips in-house back from the fab. I would think two months or so to ramp production, but that's just a guess on my part. Taking into account the Chinese New Year though, I would personally think March would be a stretch...but...I imagine what they're talking about is the A15 silicon what may be used in the 'x2850' rather than the 'x2800'. It's possible they already have, or have had for awhile, A12 in production mode gearing up for the March launch. If nothing else, and this final silicon is not used for the March launch, we could see the A15 products shortly afterwards (April/May?)
 
Ailuros said:
It's approximately 6 weeks from now to the second week of March. How much more do they need for mass production? (granted there are also the 2 weeks halt for the chinese new year, but I don't think it'll change anything dramatically in the end).
Click to expand...
Isn't the Inq article saying that they are already in mass production? Or at least mass production of the chips - production of the actual cards is presumably just about to begin. Surely that's what they mean by "production silicon" - the first of the mass-produced chips (as opposed to samples).
 
34cm, 2x6-pin (no 8-pin connector) XTX:

http://www.vr-zone.com/index.php?i=4529

That's almost 13½ inches long...That's gotta be the OEM version for high-end vendors whom sell in a large chassis (Alienware ect). I think we already went through the case debacle (did I spell it right that time?) about only certain cases have 14 inches front-to-back.
 
turtle said:
Well, we've usually seen 3 months between tape-out and availability. I wouldn't know how much time is subtracted if they already have chips in-house back from the fab. I would think two months or so to ramp production, but that's just a guess on my part. Taking into account the Chinese New Year though, I would personally think March would be a stretch...but...I imagine what they're talking about is the A15 silicon what may be used in the 'x2850' rather than the 'x2800'. It's possible they already have, or have had for awhile, A12 in production mode gearing up for the March launch. If nothing else, and this final silicon is not used for the March launch, we could see the A15 products shortly afterwards (April/May?)
Click to expand...

Actually you can change the metal layer after production has started, as the metal layer added late in production pipe. So even if A15 did not exist when the production started, the final chips can end up as stepping - there are only a 3 month delay for silicon changes (and A12 and A15 is the same silicon).
 
Tim said:
Actually you can change the metal layer after production has started, as the metal layer added late in production pipe. So even if A15 did not exist when the production started, the final chips can end up as stepping - there are only a 3 month delay for silicon changes (and A12 and A15 is the same silicon).
Click to expand...

Thanks for that information.

The question then is, would ATi change it mid-production? The chips already made could go into lower-end chips then, in theory, but that also begs the question:

Did A12 revision chips run so hot that only OEM's will initially get the XTX version (with the big-ass cooler, ~13'') while retail gets perhaps the XT/XL/(GTO) running at slower speeds on the 9'' PCB until A15 comes out and they can ramp to the speeds while using a smaller cooler (9'' pcb?).

At least seems a possibility, if enough A15's are not produced. If nothing else, whatever uses A15 at launch could be probably be rare-er, if not delayed past the initial launch if at least SOME parts are using A12. Remember the X1800XL availability, and the X1800XT not trickling into the market until a little later? It seems that could happen again.
 
turtle said:
Thanks for that information.

The question then is, would ATi change it mid-production? The chips already made could go into lower-end chips then, in theory, but that also begs the question:

Did A12 revision chips run so hot that only OEM's will initially get the XTX version (with the big-ass cooler, ~13'') while retail gets perhaps the XT/XL/(GTO) running at slower speeds on the 9'' PCB until A15 comes out and they can ramp to the speeds while using a smaller cooler (9'' pcb?).

At least seems a possibility, if enough A15's are not produced. If nothing else, whatever uses A15 at launch could be probably be rare-er, if not delayed past the initial launch if at least SOME parts are using A12. Remember the X1800XL availability, and the X1800XT not trickling into the market until a little later? It seems that could happen again.
Click to expand...

Perhaps it's as simple as , Ati started production on the A12 spin . This met the goals for OEM cards. They built up a sizable quanity for launch while switching to A15 for the retail launch. That way they could serve both sectors without under supplying both.
 
True enough. Obviously, that is the best case scenario and certainly possible. It'd be nice to see all 9'' (retail) cards using the A15 revision, just hope it doesn't delay the darn things longer.

On the other hand, if I were a high-end OEM, I'd be pissed with getting the ungodly hot/not clocking as well, huge-fanned, full-case sized pcb w/ beta-chip monstrosity to pitch in my machines. (Yeah, that was a bit over-the-top...)

I suppose we shall see how it unfolds.
 
Status
Not open for further replies.

Similar threads

A
The Official G84/G86 Rumours & Speculation Thread
22 23 24
Replies
460
Views
73K
Twinkie
T
Back
Top