GeForce 9900 GTX & GTS Slated For July Launch

[Scimitar]

If the gt200 shaders are really 50% faster clock-per-clock than the shaders on g80/g92 and gt200 has the same shaderclk as g80 it would have 181% more shaderpower.

I wonder if they are using the same compression algorithim they used on the 9600GT to get the 50% better performance, or if it is something else.

[DilTech]

Don't remember the results of the HD3870x2?

Ok, I exaggerated by 17%.

Even at the same power percentage increase, you're still looking at 255-256 for the 4870x2, so my point still stands.

[LordEC911]

Ok, I exaggerated by 17%.

Even at the same power percentage increase, you're still looking at 255-256 for the 4870x2, so my point still stands.

So you think AMD/ATi isn't improving their power saving techniques?
I remember an interesting Q/A from sometime ago talking specifically about it.

[AliG]

we're talking load here, not idle, of course the 4870x2 will consume less in idle, unless nvidia can shut down parts of the die (which I believe so far only k10 can do - and only to some degree), the single running rv770 gpu will consume far less power at idle. Thing is, ati has to be careful about when the second gpu kicks in, that's been a big problem with past dual gpu cards in power saving mode (which you can disable) as sometimes you end up with only 1 gpu running when you should have 2 and you get lower fps.

[GoldenTiger]

If the 280 gtx performs as good as it sounds, I will definately be buying it depending on whether the gtx or hd4870x2 performs the best. Do you think my corsait 520hx psu will handle it? I'm running an hd2900xt and a quad which are all doing good. Buying a new PSU is something I'm not tempted at doing.

I'm banking on mine handling it (520HX Corsair). Time to sell off the SLI 8800GT 512s . I am kind of concerned still about whether a PCI-E 2.0 slot will be needed to get the most out of this though I'm sure it's fine, as I don't have one .

[Monkeywoman]

The upcoming GeForce GTX 260/280 GPUs are based on the GT200 (NV60) core and will be built using a 65 nm manufacturing process at TSMC. However, we heard that not many chips will actually fit on a 300 mm wafer, since Nvidia has come up with a huge die measuring 24 x 24 mm, resulting in a die area size of 576 mm2. This area is almost 100 mm2 larger than Nvidia’s previous 90 nm high-end GPU (G80) and a consequence of 16 processing blocks (G80 came with nine blocks, eight were enabled for GTX models, six for GTS versions) and a new 512-bit memory controller, which replaces the old 384-bit model (the GTS260 will integrate a 448-bit version). The current G92 or GeForce 8800GT/8800GTS512/9800GTX/9800GX2 CPUs are built in a 65 nm process and end up at a die size of 330 mm2. With its new GPU generation, Nvidia is going to continue on the safe route and plan with enough spare transistors for 240 shader units (actually, 240FP+240MADD). The same will be the case with the GeForce GTX 280 and 260. The GPU will have 15 processing units (240 shader processors) available on the GTX280, while the GTX260 will come with 12 units for a grand total of 192 shader processors. Our sources state that the manufacturing cost of the GT200 die is somewhere between $100 to $110 per piece.

source; http://www.tgdaily.com/content/view/37554/135/

[MpG]

Not unthinkable, but the 250W seems a little odd.

When the 8800GTX hit the scene, people were claiming it pulled 225W, since that was the sum of the max wattages of the PCI-E slot+2 PCI-E connectors. They were also claiming that R600 needed 250W, since "apparently" ATI didn't think 2 6-pin connectors were adequate.

Personally, don't care if this thing wants 500W, as long as it uses all that power for performance. Mind, if it DOES use that kind of power, the PSU industry might just cream its collective pants.

[Lightning_Rider]

So I guess that means it's going to be insanely hot? Man I was hoping it wouldn't be such a barbecue grill this time around.

I'm excited to see benchmarks though... wonder when the NDA is lifted?

Also... shouldn't this thread's title be changed to 280GTX and June 18 now?

[LordEC911]

The upcoming GeForce GTX 260/280 GPUs are based on the GT200 (NV60) core and will be built using a 65 nm manufacturing process at TSMC. However, we heard that not many chips will actually fit on a 300 mm wafer, since Nvidia has come up with a huge die measuring 24 x 24 mm, resulting in a die area size of 576 mm2. This area is almost 100 mm2 larger than Nvidia’s previous 90 nm high-end GPU (G80) and a consequence of 16 processing blocks (G80 came with nine blocks, eight were enabled for GTX models, six for GTS versions) and a new 512-bit memory controller, which replaces the old 384-bit model (the GTS260 will integrate a 448-bit version). The current G92 or GeForce 8800GT/8800GTS512/9800GTX/9800GX2 CPUs are built in a 65 nm process and end up at a die size of 330 mm2. With its new GPU generation, Nvidia is going to continue on the safe route and plan with enough spare transistors for 240 shader units (actually, 240FP+240MADD). The same will be the case with the GeForce GTX 280 and 260. The GPU will have 15 processing units (240 shader processors) available on the GTX280, while the GTX260 will come with 12 units for a grand total of 192 shader processors. Our sources state that the manufacturing cost of the GT200 die is somewhere between $100 to $110 per piece.

source; http://www.tgdaily.com/content/view/37554/135/

LoL... Someone was right...
Interesting that they are saying 16 and 12 shader clusters. Weren't we thinking 10 and 8 clusters?
Also, what is with the transistor count? 900m-1.1b?
I'm still betting on 1.3-1.4b.

Not unthinkable, but the 250W seems a little odd.

When the 8800GTX hit the scene, people were claiming it pulled 225W, since that was the sum of the max wattages of the PCI-E slot+2 PCI-E connectors. They were also claiming that R600 needed 250W, since "apparently" ATI didn't think 2 6-pin connectors were adequate.

Personally, don't care if this thing wants 500W, as long as it uses all that power for performance. Mind, if it DOES use that kind of power, the PSU industry might just cream its collective pants.

Ummm... certain people were stating that R600 was pulling 300w per card...
TDP is 250w, that doesn't mean it is going to pull 250w at load.

[Tim]

If those specs are true, I don't think ATI stands much of a chance to begin with.

[fireice2]

ATI may not grab the performance crown but if they will price their card right, they will capture a different market.

[Yakyb]

Also... shouldn't this thread's title be changed to 280GTX and June 18 now?

arrrgh im buying a house on the 17th

[Beefy22]

The first 2 models from the GeForce 9900 series will be 9900 GTX and 9900 GTS.It seems Nvidia has decided to launch them as early as July to counter RV770 and steal the performance crown away. Nvidia has earlier confirmed GT200 existence and revealed that the GPU has about 1B transistors. Our sources are suggesting a 512-bit memory interface with GDDR3 memories for GeForce 9900 GTX using P651 PCB and a 448-bit memory interface for the GeForce 9900 GTS.

Anyone have a clue what PCB P651 costs per piece or 1GB of GDDR3? Figured it would be interesting to try and guesstimate the total manufacturing cost of GT200, since we have a rough estimate on GT200's die cost it shouldn't be too difficult.

http://www.tgdaily.com/html_tmp/cont...37554-135.html

Our sources state that the manufacturing cost of the GT200 die is somewhere between $100 to $110 per piece.

[clonez]

i would expect max. 50$ for the pcb, 30-40$ for the rams (small bin), 15-20$ for the cooler and an additional 20$ for nvio chip, caps and VRs

producing a card like that shouldnt cost more than 225$ (gpu is damn expencive, costs nearly as much as the whole 9800 gtx card lol)

[Shintai]

i would expect max. 50$ for the pcb, 30-40$ for the rams (small bin), 15-20$ for the cooler and an additional 20$ for nvio chip, caps and VRs

producing a card like that shouldnt cost more than 225$ (gpu is damn expencive, costs nearly as much as the whole 9800 gtx card lol)

GT200 GPU for nvidia would cost below 100$. Its a high profit card. If it cost 500$ expect nVidia to put atleast 200$ in their pocket.

[Calmatory]

GT200 GPU for nvidia would cost below 100$.

According to what?

[NH|Delph1]

GT200 GPU for nvidia would cost below 100$.

Really? I'm hearing +$100. The chips is just too damn big and the wafers too damn small

//Andreas

[Tim]

According to what?

According to Shintai.

He magically knows how much the Die costs.

[Shintai]

According to Shintai.

He magically knows how much the Die costs.

You can pretty much put 2 and 2 together when you know the price of about everything else.

[DilTech]

Tom's has the 9900gtx apparently...

http://www.tomshardware.co.uk/hdmi-d...ews-28242.html
Look at the last image on the page, near the bottom.

[Luka_Aveiro]

Tom's has the 9900gtx apparently...

http://www.tomshardware.co.uk/hdmi-d...ews-28242.html
Look at the last image on the page, near the bottom.

Is it possible the 9900GTX is in fact a 9800GTX photoshoped???

[DilTech]

Is it possible the 9900GTX is in fact a 9800GTX photoshoped???

It's very possible, I just saw it posted on EOCF and thought I'd pass on the info here.

[cstkl1]

I'm banking on mine handling it (520HX Corsair). Time to sell off the SLI 8800GT 512s . I am kind of concerned still about whether a PCI-E 2.0 slot will be needed to get the most out of this though I'm sure it's fine, as I don't have one .

eh and what are u going do about the 8 pin pcie.. do some voodoo magic??
kekeke

[Nuker_]

eh and what are u going do about the 8 pin pcie.. do some voodoo magic??
kekeke

What? HX series has multiple 8-pin pcie (6+2 pin).

[cstkl1]

What? HX series has multiple 8-pin pcie (6+2 pin).

damn .. just read the review.. didnt know that..
always assumed only the high wattage had it other than pcp..

ok my bad..

one thing guys if this card is as long as it is.. forget p5q deluxe
having so much trouble with it and 9800gx2..it closes of one sata port..