The Official RV630/RV610 Rumours & Speculation Thread

[aeryon]

It's a bit fishy marketing, if true like it seems it is;
Sure, they never said anywhere that UVD would be part of the full line-up, but they surely did give the impression it would be

Nobody will argue against the fact that AMD/ATI marketing sux*d badly hard since months (DX10 family launch joke, perfect R600 drivers fiasco, obscure UVD informations on R600, renaming RV515 as HD2300 and so on)

I think they must fire a bunch of guys in the marcomm dept as they create more confusion and bad publicity than nothing else...

And you know the last scandal ? What about a Radeon Mobility X2500 chip which is nothing else than an old rebadged RV530 SE ? This chip (X2500) is not listed in AMD/ATI website :
http://ati.amd.com/products/mobile.html
and some laptops start to use it:
ASUS F3SE notebook
Some customers already bought this notebook thinking about getting a DX10 GPU (2x00 series) inside and of course they've been fooled. Now they start to complaint in French forums and finally they will hate AMD/ATI. what a brilliant plan from AMD/ATI...

These dirty marketing practices drive me mad

 

[Speccy]

renaming RV515 as HD2300

Did RV515 contain UVD previously? I don't recall it.

These dirty marketing practices drive me mad

Isn't product rebranding fairly frequent? I seem to remember an NV44 (7100 GS) rebrand not so long ago. NV17 also seemed to be rebranded multiple times, and didn't have the same functionality as parts within its apparent series.

 

[Geeforcer]

So what exactly is HD 2300? One thing for sure, it does not fit into ATI's own description of 2000 series.

 

[Topman]

I don't know... no HD 2600XT GDDR4 in Sapphire's Press Release?
(ATi Nº1 Partner)

Maybe HD 2600XT GDDR4 =~ HD 2900 XTX -> impracticable(costs/performance)?
Some partners will launch this cards... like Gecube...

But I'm confident on HD 2600XT GDDR3...

http://www.hexus.net/content/item.php?item=8789

bye

 

[aeryon]

Did RV515 contain UVD previously? I don't recall it.


Isn't product rebranding fairly frequent? I seem to remember an NV44 (7100 GS) rebrand not so long ago. NV17 also seemed to be rebranded multiple times, and didn't have the same functionality as parts within its apparent series.

1/ no UVD engine on RV515. (by UVD I mean the full VC1/H264 engine found on RV630/610. see my posted graph above)

2/ frequent or not isn't the problem. In consumer point of view, rebranding is very bad practice and when we see some cases of it, we must denounce it.

So what exactly is HD 2300? One thing for sure, it does not fit into ATI's own description of 2000 series.

last bottom product of the page. ATI Mobility Radeonâ„¢ HD 2300

 

[Geeforcer]

last bottom product of the page. ATI Mobility Radeonâ„¢ HD 2300

Yeah, I saw that. I just don't understand what it is, exactly, if it's not RV515. It apparently has 10 million more transistors than X1300, no DX10 support,(contrary to ATI's description of "2000-series"), is made at 90nm and appernlty has UVD. I wonder if someone from ATI could shed some light on this.

 

[AnarchX]

Yeah, I saw that. I just don't understand what it is, exactly, if it's not RV515. It apparently has 10 million more transistors than X1300, no DX10 support,(contrary to ATI's description of "2000-series"), is made at 90nm and appernlty has UVD. I wonder if someone from ATI could shed some light on this.

RV550 = RV515/RV516 + Xilleon video processor

 

[HAL]

http://www.fudzilla.com/index.php?option=com_content&task=view&id=1122&Itemid=1

 

[vertex_shader]

I don't know... no HD 2600XT GDDR4 in Sapphire's Press Release?
(ATi Nº1 Partner)

Maybe HD 2600XT GDDR4 =~ HD 2900 XTX -> impracticable(costs/performance)?
Some partners will launch this cards... like Gecube...

But I'm confident on HD 2600XT GDDR3...

http://www.hexus.net/content/item.php?item=8789

bye

This card has Valve black box voucher too :smile:

 

[Speccy]

1/ no UVD engine on RV515. (by UVD I mean the full VC1/H264 engine found on RV630/610. see my posted graph above)

Click the link in the post - UVD is clearly listed (although its not for X 2300). Seems like RV610 and RV630 are not the only ones with UVD and this isn't an RV515.

 

[Arun]

and this isn't an RV515.

RV550?

 

[Razor1]

http://www.fudzilla.com/index.php?option=com_content&task=view&id=1122&Itemid=1

if that is true, gotta wonder if AMD is having problems with the 65nm process as well, but then what is causing TSMC to be capacity constrained?

 

[Arun]

Razor1: Well, it is not unthinkable that the child is refusing the lessons of his non-biological father. I'm fairly convinced that nothing out of the ordinary will happen and RV610/RV630 will get good market traction in all segments though, but we'll see.

P.S.: Do I win a 'most cryptic post of the month' award?

 

[neliz]

Yeah, I saw that. I just don't understand what it is, exactly, if it's not RV515. It apparently has 10 million more transistors than X1300, no DX10 support,(contrary to ATI's description of "2000-series"), is made at 90nm and appernlty has UVD. I wonder if someone from ATI could shed some light on this.

I haven't seen HD2300 up to then, but it's probably a typo as the cards should be X2100 and X2300 (for dx9) and HD2400 etc. for DX10. renaming schmenaming.

 

[Razor1]

Razor1: Well, it is not unthinkable that the child is refusing the lessons of his non-biological father. I'm fairly convinced that nothing out of the ordinary will happen and RV610/RV630 will get good market traction in all segments though, but we'll see.

P.S.: Do I win a 'most cryptic post of the month' award?

LOL definitly don't disagree with that .

I don't know much about % of volume production coming out of TSMC, but a few years ago, around 30% was GPU's (dont' know if this is number of chips coming out or % of capacity), it might have changed since then but what it sounds like to me AMD might be having some problems with yeilds (if this "report" is accurate), I wouldn't say it is the same leakage problems as the r600 though.

 

[Zaphod]

I haven't seen HD2300 up to then, but it's probably a typo as the cards should be X2100 and X2300 (for dx9) and HD2400 etc. for DX10. renaming schmenaming.

In the mobile space, X2300 and HD 2300 seems to be two distinctly separate things. The former being morphed from the Mobility X1400 (M64), but there's also an upcoming Mobility Radeon HD 2300 (M71). According to the AMD website, Linux drivers, and the PCI-SIG lists: the only spec difference I can find is the mention of "UVD - Full H.264 and VC-1 decoder" for the latter.

At this point, there doesn't seem to be any indication that they'll launch products based on the desktop RV550 counterpart, though. Gone the way of the Dodo?

 

[Geo]

if that is true, gotta wonder if AMD is having problems with the 65nm process as well, but then what is causing TSMC to be capacity constrained?

That "greater than 100%" bit out of digitimes or wherever interests me too. "100%" has a whole bunch of assumptions built into it, of course, re number of shifts, routine maintenance downtime, etc. Since Fab equipment is incredibly expensive, we may (it seems to me) reliably assume that TSMC works hard to maximise the amount of time its in use already.

One of the assumptions built in, and the easiest way, I bet, to get "greater than 100% usage" is by running particularly large orders. One of the assumptions is going to be how much time you lose transitioning the fab from "GPU A" to "other customer B's ASIC". Planners will have built in assumptions about average order size and time to switch from one to the next in defining "100%". Just to pluck some numbers out of the air, if the planners are assuming they'd have 100 different ASIC orders of 100k each but in reality they have 1 order of 10M. . .well, then there's going to be a significant time savings there that's going to shoot you past "100%" in that timeframe.

 

[Jawed]

What I'd like to know is how "TSMC's 65nm process is running at full tilt" becomes "AMD is having problems with 65nm"?

What about "AMD is buying up all of TSMC's available 65nm capacity, and then some"?

Jawed

 

[Razor1]

well if AMD is hogging all 65nm capacity, where are these chips/cards? Why isn't anyone selling these cards? Keeping a stock pile of hundreds of thousands of chips/cards, and not getting them to partners, OEM's, system vendors (haven't seen a single online site where you can buy these cards yet) and what not is not a smart business plan. When did AMD state they are selling millions of these cards, March, So I don't think AMD's ordering huge amounts of chips is the cause of the supply constraint, no matter how you look it. How much do they need to launch these cards? maybe 500k? When did they start production on these chips? Lets say March since AMD seemed confident they were getting millions of chips sold. And if these cards don't show up in June, and come out in July, the reason for the supply contraint really falls through.

 

[Jawed]

So, you're saying that everyone other than AMD is hogging 65nm and there's nothing left for AMD to work with.

Jawed