GeForce 9900 GTX & GTS Slated For July Launch

[Mortal]

I think I'm gonna go with the GTX 260[/I])

I'm in the same boat. My 320MB GTS has served me well, but its time to move on up.

[Hornet331]

dudes read the name of the person your are quoting , this is "Dynasty" that guy doesn't need a ss to back up what he is saying , if you don't want to believe it just don't read it , you don't need high 3D mark score to make a good card

sorry, but i prefer logic over blind obedience, this nubers dont add up and make no sense, a 280gtx slower then a 8800gts?
With everything shaders doubled, more tmus more rops etc on the 280gtx , that makes just no sense.

[kemo]

Well kingpin is no.1 in hwbot but they stil ask him for an SS for verification

Not under NDA

sorry, but i prefer logic over blind obedience, this nubers dont add up and make no sense, a 280gtx slower then a 8800gts?
With everything shaders doubled, more tmus more rops etc on the 280gtx , that makes just no sense.

As far as i know with a 3GHZ Yorkie
GTS 512 13.7K

so it can be considered a bit faster for 3D mark 06 which is pointless and extremely limited Not to mention that Dynasty didn't say anything about filtering and so , anyway this a pointless argue , enough spamming , i will wait and see

[The Nemesis]

3D 06 1280x1024

GTX 280 - 14700

Card is running @ 605 core /1110 mem /1300 shader speeds.

QX9650 C0 @ 3100mhz
eVGA 790i ultra
Super Talent DDR3 @ 1600 CL 7

649.00 USD MSRP sounds about right may even go as high as 700.00 first
week or so.

This must've been on vista...even so

Q9450 @ 3.8GHz 8800GTX 648/1026

Q9450 @ 3.7GHz 8800GTX 648/1026

[zerazax]

Again, you CANNOT compare cards across different systems via 3dMark scores... CPU affects the 3dMark06 score way more than the actual card itself and who knows how driver performance atm is

[The Nemesis]

Again, you CANNOT compare cards across different systems via 3dMark scores... CPU affects the 3dMark06 score way more than the actual card itself and who knows how driver performance atm is

I agree, which is why I posted two results where the difference is 100MHz on the cpu clock. It seems 3DMark06 will give an additional 100Pts for each 100MHz on a Quadcore clock. My comparison I thought was showing the potential. If I was home, I would just clock my cpu to 3GHz and core to 600 but then my score would barely break 13,000. I think 14,700 with a single card with no overclocking of card, cpu, or system is pretty good. I think most were just expecting something spectacular like 18,000 or so.

[serialkilla277]

i want to see some 05 03 01 results ...

[Nedjo]

i want to see some 05 03 01 results ...

ROFL... I presume that for OpenGL performance you would like Quake ][ timedemo?

[jas420221]

As far as i know with a 3GHZ Yorkie
GTS 512 13.7K

Stock clocks on everything (see sig) I think I got 14k. I barely broke 17k with settings in sig.

Unless something is holding that card back...that score is a damn shame.

[Luka_Aveiro]

I remember seeing some charts on toms hardware vga comparison, a 8600GTS beating a 8800Ultra @ 1024x768 in some game, due to higher frequency clocks.

I am not surprised the card scores this low, I quite believe this card shines with filters, so a 1280x1024 test without filters doesn't seem to me the best playground for this card.

Compare it with riding a car using rain tires on gravel.

[YukonTrooper]

ROFL... I presume that for OpenGL performance you would like Quake ][ timedemo?

Hey man, look at his avatar. Maybe he wants to play WoW at 10,000 FPS.

[adamsleath]

I quite believe this card shines with filters

yeah that's the deal for me; i want 8xQ minimum , so that's what i'll be looking for in reviews, all the candy in game.
the only reason i have a gcard is for eye candy.

[msgclb]

You have to love FUD.....

He still thinks GT200 has 8 clusters, meanwhile a die shot clearly shows it has 10. The GTX 260 has 8/10 enabled.... how does he expect the card to have 7/8 clusters enabled yet have 192 SPs?

Consider the following statement:

Nvidia’s GeForce 9800 GTX features 16 ROPs. The ROPs are split into four clusters – each is able to process four pixels per clock and is also connected to a 64-bit memory interface. This means that there’s a 256-bit memory interface on the GPU and it connects out to eight 64MB GDDR3 DRAMs, making a total of 512MB of video memory.

If that's true and I believe it is then I'm going to use some reverse logic to theorize the following:

Nvidia’s GeForce GTX280 features 32 ROPs. The ROPs are probably split into eight (dividing 32 by 4) clusters – each is able to process four pixels per clock and is also connected to a 64-bit memory interface. This means that there’s a 512-bit memory interface on the GPU and it connects out to sixteen 64MB GDDR3 DRAMs, making a total of 1GB of video memory.

Nvidia’s GeForce GTX260 features 28 ROPs. The ROPs are probably split into seven (dividing 28 by 4) clusters – each is able to process four pixels per clock and is also connected to a 64-bit memory interface. This means that there’s a 448-bit memory interface on the GPU and it connects out to fourteen 64MB GDDR3 DRAMs, making a total of 896MB of video memory.

I know that I've seen how many memory chips there are on these boards somewhere but I didn't take to time to see it is 16 and 14 as I stated above.

Are we talking about the same clusters?

[Extelleron]

Consider the following statement:


If that's true and I believe it is then I'm going to use some reverse logic to theorize the following:

Nvidia’s GeForce GTX280 features 32 ROPs. The ROPs are probably split into eight (dividing 32 by 4) clusters – each is able to process four pixels per clock and is also connected to a 64-bit memory interface. This means that there’s a 512-bit memory interface on the GPU and it connects out to sixteen 64MB GDDR3 DRAMs, making a total of 1GB of video memory.

Nvidia’s GeForce GTX260 features 28 ROPs. The ROPs are probably split into seven (dividing 28 by 4) clusters – each is able to process four pixels per clock and is also connected to a 64-bit memory interface. This means that there’s a 448-bit memory interface on the GPU and it connects out to fourteen 64MB GDDR3 DRAMs, making a total of 896MB of video memory.

I know that I've seen how many memory chips there are on these boards somewhere but I didn't take to time to see it is 16 and 14 as I stated above.

Are we talking about the same clusters?

We're talking about different things.

I am talking about the SP clusters, of which the SPs and the TMUs are a part of. GT200 has 10 SP clusters, each with 24 SPs and 8 TMUs. The GTX 280 has all 10 enabled, so that gives it 240 SP / 80 TMU. The GTX 260 has 8 enabled, so that gives it 192 SP / 64 TMU.

You are talking about ROPs, which are seperate from the SP clusters but are tied directly to the memory bus width. In terms of ROPs, 28 out of the 32 found on GT200 are enabled in the GTX 260 model. So that does work out to 7/8. This is why you see the 448-bit bus on the GTX 260; 28 * 16 = 448, just like 32 * 16 = 512

FUD just doesn't know what he is talking about. Nobody refers to ROPs as "clusters" so I doubt he meant to refer to them.

[grimREEFER]

I think I'm gonna go with the GTX 260...(Xtreme Instincts slaped me in face.."idiot spend money go GTX 280")

im gonna be picking up a gtx260 most likely as well. unless a 9800gtx shows up for less than $200. nvidia hasnt mentioned any price cuts yet for it, but i have the feeling even if they don't do it officially, the invisible hand of the market will bring it's price down, just because its so inferior to all these new products coming out.

[GoldenTiger]

im gonna be picking up a gtx260 most likely as well. unless a 9800gtx shows up for less than $200. nvidia hasnt mentioned any price cuts yet for it, but i have the feeling even if they don't do it officially, the invisible hand of the market will bring it's price down, just because its so inferior to all these new products coming out.

I keep thinking of splurging on a 280, but it's just sooooooo expensive for not much gain price/performance-wise that I am leaning more toward the 260 at this point myself, as you are.

[comc49]

so is 9900gtx gtx280? or is it 55nm of 9800gtx???

[YukonTrooper]

so is 9900gtx gtx280? or is it 55nm of 9800gtx???

GTX280 is GTX280. 9900GTX is a G92 core 55nm refresh.

[Hornet331]

http://publish.it168.com/2008/0611/20080611033801.shtml

new benches... lol now the 280GTX only reaches 12,5k in 3dmark 06....

[PurpleMonkey910]

http://publish.it168.com/2008/0611/20080611033801.shtml

new benches... lol now the 280GTX only reaches 12,5k in 3dmark 06....

so it reaches less less than the 8800GTX i had & the 9800GX2 i have? I think they may be making that up

[Philip_J_Fry]

probably driver issue ... even if those are true that is .. seems like nv is focused on DX10 game this gen

[The Nemesis]

Looking at the chart confirms whats already been said. The card performs much better than current offerings @ higher resolutions and eye candy cranked up. Still if the price of the current hardware drops significantly for something like the 9800GX2 I'd buy a second before shelling out $700 on the GTX280

[msgclb]

Do you believe this? If true, I'll go on a diet for the next year!

[The Nemesis]

Do you believe this? If true, I'll go on a diet for the next year!

That'd be awesome! It probably will happen too. 2 months after its release

[jpcastro]

attention attention, inquirer news

GT200 scores revealed

THANKS TO NVIDIA'S shutting us out, we are not handcuffed about the GT200 numbers, so here they are. Prepare to be underwhelmed, Nvidia botched this one badly.

Since you probably care only about the numbers, lets start out with them. All 3DMark scores are rounded to the nearest 250, frame rates to the nearest .25FPS. The drivers are a bit old, but not that old, and the CPU is an Intel QX9650 @ 3.0GHz on the broken OS, 32-bit, SP1.

http://www.theinquirer.net/gb/inquir...cores-revealed