PDA

View Full Version : GT200 release confirmed?


aVaLaNcHe
04-01-2008, 04:06 PM
From VR-Zone (http://sg.vr-zone.com/articles/NVIDIA_GT200_:_GeForce_9900_GTX_&_9900_GT_In_Q3/5684.html):
No it is not some April Fool jokes. VR-Zone has learned that there will be two high end parts from Nvidia to replace the current GeForce 9800 GX2 and 9800 GTX in Q3. These two parts are GeForce 9900 GTX and GeForce 9900 GT based on the new GT200 architecture. GeForce 9800 GX2 and GTX will have a short lifespan since they will be replaced in a couple of months time. Nvidia has positioned GeForce 9900 GTX to replace 9800 GX2 and this gave a clue that a single GT200 card will actually outperform a dual G92 card. Likewise for GeForce 9900 GT, outperforming and replacing 9800 GTX. CJ revealed to us that Nvidia will have a hard time fitting GT200 onto the PCB since the GPU is gonna be large and very power hungry so we can probably assume it is still 65nm based. The PCB for GT200 will be P651.

From Expreview (http://en.expreview.com/2008/03/31/gt200-to-become-9900gtx-and-9900gx2-launch-date-set-to-july/#more-317):
According to our sources, GT200’s final product name and launch day have been set. Two NVIDIA GeForce card will use the core, they are 9900GX2 and 9900GTX.

It is similar to today’s NV tactic: 9900GX2 just like todays 9800GX2, it is a two GPU product, and 9900GTX is a conventional single GPU product. Source mentions that both cards will be announce in July. In previous rumors, words have it that NV will improve G9X core by rolling out GT200. But detailed information still unknown.


Will be interesting to see the final details and numbers. If GT200 is indeed faster than dual G92, I will be upgrading in July :D

Sunny
04-01-2008, 05:54 PM
Very Interesting stuff. I would definately like them to surpass those cards as they are not so much of an improvement over the 8800 GTX/Ultra. If they proved very powerful I might consider upgrading too.

Casecutter
04-01-2008, 07:59 PM
It sounds as though Nvidia is sending up a trial balloon. They are looking to see if there’s any reason to press this to market or keep it in R&D to sort things out further. If there’s enough chatter that the übber-enthusiasts are ready and willing pay for an actual “new” architecture, they’ll skip any more development and get it to Fab. Even if such cards forgo being less power hungry, they’ll offer these up for those still salivating for something more than a GX2 offered.

This also has much to do with what and when ATI has something showing. As Nvidia would have these higher-end cards to steal thunder and reviewer time from (hopefully) a new mainstream offering (less juicy) positioned for price and performance (value), something that isn’t so much coming across by this bullshit… :p

Radiator
04-01-2008, 09:28 PM
Looking forward to it... but if it's going to be so power hungry , why not use 55nm ?

Sunny
04-01-2008, 09:42 PM
Yeah they should be definately looking to decreasing the die size. I know my 8800 GTX's get pretty hot on air and I wish they had a smaller GPU size.

OH man speaking about 55nm. I bought the HD 3870 Toxic edition by saphire and its soo cold on air its retarded. It was idling yesterday at 38C on air. What the HECK! I was amazed. My 8800 GTX's don't even reach that on water.

Casecutter
04-01-2008, 10:51 PM
A die shrink takes more than engineering, money, and good Fab house to produce chips in volume. Nvidia got slapped silly the last time (90nm -65nm didn’t go to well) when they ramped up production on G92… I think they’ll wait to make certain the new architecture is able to undergo a shrink. They believe we’ll pay-up for whatever they dish-up, and unfortunately I think they have the market pegged. :rolleyes:

Radiator
04-02-2008, 02:03 PM
Yeah they should be definately looking to decreasing the die size. I know my 8800 GTX's get pretty hot on air and I wish they had a smaller GPU size.

OH man speaking about 55nm. I bought the HD 3870 Toxic edition by saphire and its soo cold on air its retarded. It was idling yesterday at 38C on air. What the HECK! I was amazed. My 8800 GTX's don't even reach that on water.

You have three 8800 GTX's ... obviously their temperatures will reach riddiculous amounts , even with water ( the cards are next to eachother , so I suppose they heat up the waterblocks a bit ) .

Anyways... I'm confident , that the R700 will be more competitive than the R600 was...
But yeah... good to get an user's confirmation that the HD 3870 is actually cool... it was a hell of a supprise for me , because as we all know , the HD 2900 XT reached temperatures other cards couldn't even reach , unless overclocked like hell .

aVaLaNcHe
04-02-2008, 05:08 PM
OH man speaking about 55nm. I bought the HD 3870 Toxic edition by saphire and its soo cold on air its retarded. It was idling yesterday at 38C on air. What the HECK! I was amazed. My 8800 GTX's don't even reach that on water.

Both of my 8800GT cards idle at 37C. I've never had a card with a G80, do they really heat up like that?

Radiator
04-02-2008, 05:15 PM
They are on 90nm process ... so yeah .
Sunny has the worst case scenario , though ... he has three of them .

Sunny
04-02-2008, 07:09 PM
Both of my 8800GT cards idle at 37C. I've never had a card with a G80, do they really heat up like that?

Yeah the GTX gets pretty hot.
This is on air and take into consideration I have the fans at 80-100%.
1 Card = idle low to mid 50's on Load about low to mid 60's
2 Cards = idle mid 50's and load is about high 60's
3 Cards = idle low 60's and load about mid to high 70's

Also just to mention my Case temps where about high 40's compared to about mid 20's now.

On water I get idle anywhere from 40-45C the cards will vary a bit and on load I have never seen it pass 54C which is not bad with water.

Sunny
04-02-2008, 07:12 PM
They are on 90nm process ... so yeah .
Sunny has the worst case scenario , though ... he has three of them .

Yeah three of them on air is sooo hot its retarded I don't see any other choice but to watercool when you have three.

I only wonder what temps I would get on the HD 3870 water cooled. I would probably cry if they got so low. ;)

Headfoot
04-02-2008, 08:10 PM
Process changes can be extremely expensive if the process isnt fully mainstream yet. The yields are never as good on the new size. ATi used 55nm because they were trailing, and it helped them catch up back into a competitive segment. :o

My cards if I dont turn the fans to 100% can idle at 60-something, but they usually hit load at 71-72, so they have a smallish range, its just really high.

Sunny
04-02-2008, 08:37 PM
My cards if I dont turn the fans to 100% can idle at 60-something, but they usually hit load at 71-72, so they have a smallish range, its just really high.

Whoa your card is clocked pretty high though. Nice job.

Radiator
04-02-2008, 08:45 PM
Yeah three of them on air is sooo hot its retarded I don't see any other choice but to watercool when you have three.

I only wonder what temps I would get on the HD 3870 water cooled. I would probably cry if they got so low. ;)

I don't think they'd go super low , because well... water is always above zero .

Sunny
04-02-2008, 11:59 PM
Probably not super low but lower then air.

-RK
04-03-2008, 01:09 AM
Get some phase change going in there! Better yet, get an 8800GTS G92 and mess around with some 2+ ghz shader clocks!

woo-hoo refrigeration! Good luck finding it though...

If you live in a humid environment forget about it, condensation can be a bitch.

Headfoot
04-04-2008, 02:30 AM
Whoa your card is clocked pretty high though. Nice job.

Thanks! I did have to fill every possible fan slot to get there, haha :o

Sunny
04-04-2008, 03:22 AM
Get some phase change going in there! Better yet, get an 8800GTS G92 and mess around with some 2+ ghz shader clocks!

woo-hoo refrigeration! Good luck finding it though...

If you live in a humid environment forget about it, condensation can be a bitch.

Yeah I was thinking about that phase change and stuff. I was thinking if maybe I could rig it to the watercooling setup so that all the things get cooled versus standard phase change just for the CPU. I think the water would freeze though. After seeing that the phase change things can get your processor anywhere from -35C to -65C and I think water freezes at 0 C so I don't know if that would work. My watercooling setup would be a big ice cube. I think:confused:

-RK
04-04-2008, 03:52 AM
Phase change WILL NOT work with water.

It basically works the same way a freezer does.

aVaLaNcHe
04-13-2008, 03:09 AM
Seems it's coming a little sooner than was mentioned.

http://www.hardforum.com/showthread.php?p=1032356741#post1032356741

Lots of links to sources.