The Official RV630/RV610 Rumours & Speculation Thread

Status
Not open for further replies.
Geo said:
If a GPU came off the line today, how long until a finished card showed up on Tier 1 OEM X's doorstep?
Click to expand...

IDK, but if a PCB had a date of manufacture, you could just subtract GPU's date of manufacture and add a week for transportation.
 
Isn't the technical word for "almost summer" actually "spring"? :LOL: You don't think NV's midrange parts were a bit later than they and their partners would have liked with Vista staring at everyone?
 
:LOL: thats true, it still is spring, drivers concern me more about OEM deals for back to school then anything else. AMD still has sometime to get them done but its getting very close. As you mentioned AMD seems likely to gain marketshare low end to midrange this quarter but I think it is mostly low end, many OEM's are still selling low end dx9 parts.
 
Hmm, we talk about 2 months as being fast (and costly) for a GPU respin, doesn't normal production run at a more leisurely rate? I don't know how long it is between wafer start and packaging for a ultra-high volume GPU, but it's looking like several months.

Plus there are rumours that OEMs are getting high priority. Didn't RV505 spend a while being installed in OEM computers before it was available as a discrete card purchase? Wasn't it something of a far east phenomenon?...

The timing of this capacity announcement and the timing of the AMD/TSMC partnership on 65nm announcement jointly give a rather large hint that a lot of AMD 65nm parts are in full production. Not "... gotta wonder if AMD is having problems ...".

Jawed
 
TSMC didn't have contraint problems when ATI had both Dell and HP deals, with the x800's and x600's why would they have problems now? And those chips were all on the same process.

Its not because of one of their clients that TSMC will have contraint problems. There has to be an increase from many of their clients or yeilds of the process is lower then expected, that will end up hurting everyone that is using that process node.

I vaguely remember AMD stating .65 node was very good to them, and we will be seeing cards on them sooner then expected, what changed in the interm?
 
Last edited by a moderator:
Geo said:
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".
Click to expand...
The transition time between products is essentially zero: wafers lots are moved around from one machine to the next, tagged with bar codes. Machines get automatically configured accordingly.
I doubt that a fab like TSMC in which hundreds of products from different customers are produced functions differently.

Maybe running at more than 100% means they have to man the fab with more than their usual amount of employees or reduce the amount of maintenance downtimes?
 
Not sure if this info is still relevent but..


http://www.dataweek.co.za/article.aspx?pklArticleId=4064&pklIssueId=583&pklCategoryId=31
ATI is currently one of TSMC's largest customers, but AMD has a growing relationship with Chartered Semiconductor, which is also part of the Common Platform manufacturing group with IBM and Samsung.​



http://www.theinquirer.net/default.aspx?article=32833
UMC is currently offering full production 90nm on 300mm wafers at two Fabs (Fab 12a and UMCi). It is also planning to begin 80nm production very shortly. UMC’s 65nm process looks to be some months behind TSMC and, for now, UMC will really be focusing on selling production space for both 90nm and 80nm.

Having a second option for producing complex GPUs is something that both ATI and Nvidia are quite happy with. This tends to keep the primary production partner honest and competitive (in this case TSMC). It also puts pricing pressure on TSMC to offer more value for what production lines they have.​


Is UMC still behind in the 65nm race? Anyone know?


http://forum.beyond3d.com/showthread.php?t=39379
Looks like NVIDIA is working with TSMC on the 65nm EDRAM they recently announced, it this excert is correct.



Perhaps this should go in another forum but seems telling..
http://www.eetimes.com/news/semi/sh...V0HCJH2QSNDLPSKHSCJUNN2JVN?articleID=19201721
Arithmatica, Inc., a company that claimed it could improve silicon performance and reduce die area with its innovations in mathematical circuit blocks, has revealed that Xilinx Inc. and Nvidia Corp. are two companies that have used its CellMath libraries.
The company's chief executive has also told Silicon Strategies that Arithmatica is already applying its technology to a 65-nm manufacturing process with a secret partner​

If NVIDIA was plaining on using librarys that were migrating to 65nm, coupled with plans to hit 1 TFlop in the 4th Quarter of 2007.. Then it must mean they are going from 80nm -> 65nm, which means they need TSMC to do the processing.. Which means that if they finished 8600 way back in Q2 of 2006 then the next gen should be starting production about NOW.. Correct?
 
3vi1 said:
http://www.eetimes.com/news/semi/sh...V0HCJH2QSNDLPSKHSCJUNN2JVN?articleID=19201721
Arithmatica, Inc., a company that claimed it could improve silicon performance and reduce die area with its innovations in mathematical circuit blocks, has revealed that Xilinx Inc. and Nvidia Corp. are two companies that have used its CellMath libraries.
The company's chief executive has also told Silicon Strategies that Arithmatica is already applying its technology to a 65-nm manufacturing process with a secret partner​

If NVIDIA was plaining on using librarys that were migrating to 65nm, coupled with plans to hit 1 TFlop in the 4th Quarter of 2007.. Then it must mean they are going from 80nm -> 65nm, which means they need TSMC to do the processing.. Which means that if they finished 8600 way back in Q2 of 2006 then the next gen should be starting production about NOW.. Correct?
Click to expand...

Nvidia isn't very secretive about their use of the CellMath Libraries according to that. So, I wouldn't jump to conclusions and assume that the "secret partner" is Nvidia.

Also, TSMC is hardly going to be supply-constrained by either companies High End chips. The only lines from either ATI or Nvidia that have a possibility of requiring enough capacity to overwhelm fab capacity would be their mid-end/low-end chips.

There just isn't enough demand for high end products for that considering they make up less than 5% of the market.

Regards,
SB
 
I'd send it back and get something else if the feature was that important to me before going through all of that. Then again, I have zero interest in HD-Video at the moment - DVD's and Xvid work just fine for me . Maybe if it claimed on the box that I would get more *phat lewt* by using that card, and it turned out that I got the same amount of lewt or lower quality lewt, then :mad: yeah, I might want to sue in that case :D
 
hd2400provo9.jpg

hd2400xtid6.jpg

hd2600profx4.jpg

hd2600xtod4.jpg


HD2600XT has the black box voucher.
Looks like sapphire scared now, and forget the UVD logo :smile:
 
vertex_shader said:
HD2600XT has the black box voucher.
Looks like sapphire scared now, and forget the UVD logo :smile:
Click to expand...

Wow, finally cool boxes from Sapphire (well, the bottom two - the guy on the second one looks like a girly man despite the goutee :LOL: ). Much nicer than the naked ice man "soft porn" edition :D
 
Status
Not open for further replies.

Similar threads

S
AMD: Sea Islands R1100 (8*** series) Speculation/ Rumour Thread
69 70 71
Replies
1K
Views
374K
AnarchX
A
A
The Official G84/G86 Rumours & Speculation Thread
22 23 24
Replies
460
Views
69K
Twinkie
T
A
Beyond3D's CeBit Coverage: Official Thread
Replies
12
Views
4K
_xxx_
X
A
The R7xx Architecture Rumours & Speculation Thread
5 6 7
Replies
124
Views
30K
AlexV
AlexV
A
The Official IGP Rumours & Speculation Thread
2 3 4
Replies
64
Views
11K
Ailuros
Ailuros
Back
Top