The Official RV630/RV610 Rumours & Speculation Thread

Status
Not open for further replies.
The second board has a ES/Demonstrator written all over it. Namely, the bunched-up power regulators at a weired angle.

IIRC there are some cards out there with regulators or other components (besides the mem chips of course) in weird angles, though?
 
IIRC there are some cards out there with regulators or other components (besides the mem chips of course) in weird angles, though?

That's true (as anaqer's link shows), it's just that that part of the Rv630 board looks.... sloppy? comapred to RV570 for example. The whole design looks like "Let's take RV570 and jam up more regulators". The second board looks like a much more clean layout.

BTW, what are the 4 huge chips in the back?
 
WHy is one board longer than the other? And why are they blocking the rage theater chip? Or is it the theater chip?

IIRC the RV630 doesn't have the RT200, just the R600 boards...and perhaps only the XT/XTX at that. The chip that's hidden, I believe, is the Xilleon-based UVD decoder chip.
 
IIRC the RV630 doesn't have the RT200, just the R600 boards...and perhaps only the XT/XTX at that. The chip that's hidden, I believe, is the Xilleon-based UVD decoder chip.
Given the way UVD has been mentioned in previously leaked slides -- as a feature inherent to the entire R6 line -- I would have thought it would have been more cost effective (especially for the volume parts such as RV610) to integrate the necessary logic into the main ASIC rather than add a separate chip to the PCB in the way that the Theatre 200 chip has been used in specific AIB SKUs for VIVO functionality over the last couple years?
 
IIRC the RV630 doesn't have the RT200, just the R600 boards...and perhaps only the XT/XTX at that. The chip that's hidden, I believe, is the Xilleon-based UVD decoder chip.


I was thinking something other the RT200 as I though it would be pretty funny if their hidding the same ol chip that has been around since R300?
 
Anybody else look at the date on the picture? 2007/1/11. If ATI has had samples since Janurary, I'll be one pissed off fella if these cards are delayed.
 
Anybody else look at the date on the picture? 2007/1/11. If ATI has had samples since Janurary, I'll be one pissed off fella if these cards are delayed.

the guy who took these photos said that this's because he set up the wrong date in his dc.
 
I prefer the longer board layout as the PS block is away from the monitor outs (even though everything's digital these days). With 512Mbit density GDDR4, you get a 256bit bus...
 
Given the way UVD has been mentioned in previously leaked slides -- as a feature inherent to the entire R6 line -- I would have thought it would have been more cost effective (especially for the volume parts such as RV610) to integrate the necessary logic into the main ASIC rather than add a separate chip to the PCB in the way that the Theatre 200 chip has been used in specific AIB SKUs for VIVO functionality over the last couple years?

My understanding was that the R6xx, and RV630 would have the chip integrated and the RV610 would have the Theatre chip as a seperate chip.

US
 
Interestingly, RV530 was deployed in two GPUs, X1600XT and Pro with bandwidth of 22.1 and 10.9GB/s. So, in a sense, there's a precedent for a 2:1 ratio in bandwidth based on the "pro" board. But X1600XT was blessed with considerably more bandwidth than was useful - X1650XT, with twice the ALUs, TMUs and ROPs is upto 75% faster (typically 35%) with slightly less bandwidth than X1600XT.

That score of 2105 in 3DMk06 SM3 test is pretty much the same as a stock 8600GTS, e.g. 2029:

http://www.xbitlabs.com/articles/video/display/geforce8600gts_21.html

Jawed
According to the latest Fuad update, the HD2600XT is clocked at 800/1100. (35.2GB/sec)

The cards will be shown/given to the NDA'ed press today . .
 
Something curious though, the CeBit cards should be quite a bit newer, considering the longer card is from January, could it be possible that they actually managed to pull it through with RV560 design later on?
 
considering the longer card is from January, could it be possible that they actually managed to pull it through with RV560 design later on?
Ehr? longer card from january? as posted above, the date on the photo was wrong, the guy didn't set the dat on his camera yet. Do not believe everything you see!

Mao5 said:
the guy who took these photos said that this's because he set up the wrong date in his dc.
 
Yes, this is from Fudo, but I think it would worth another updated on the thread as it is very quite at the moment. Fudo said that the Radeon HD 2600 XT works at 800 / 2200 MHz...

Sound like Fuda believed that AMD will need A14 revision to be the final chip rather than A13, and he thought it would be overclocked well.

Edit: correct some wording...
 
Given the way UVD has been mentioned in previously leaked slides -- as a feature inherent to the entire R6 line -- I would have thought it would have been more cost effective (especially for the volume parts such as RV610) to integrate the necessary logic into the main ASIC rather than add a separate chip to the PCB in the way that the Theatre 200 chip has been used in specific AIB SKUs for VIVO functionality over the last couple years?

Very possible uvd might be through the main ASIC. I got that impression from some of the information too, but some leads credence to a separate chip, at least on rv610 and/or rv630, as perhaps they still do not possess the ability to do 1080p video like the generation before them. From the naked shots of R600, the only other chip visible is the RT200, but it's clearly not the same shape as that chip. RT200 is square, and that's rectangle. Also, AFAIK the RT200 will only be on higher-end cards as it's essentially giving you the VIVO functions of the now-defunct AIW as gravy. Granted, Xilleon (in it's known form) doesn't exactly fit the bill either. I have no idea.

11763345901ZYw5Q8rmD_1_14_l.jpg

theater200_sml.jpg

questionmarked4.jpg

8HfIE_in_det_04.jpg
 
Last edited by a moderator:
Status
Not open for further replies.
Back
Top