GeForce 9-series lineup revealed

[NoobyDoo]

DigiTimes : Nvidia GeForce 9-series lineup revealed

At the top-end of the 9-series cards will sit the dual-GPU GeForce 9800 GX2 which will launch on March 11 this year. This will be followed by the GeForce 9800 GTX and 9800 GT between the end of March and beginning of April.

For the mid-range Nvidia will offer the GeForce 9600 GS in May, according the sources.

In June, the company will launch the value-level GeForce 9500 GT based on the G96 graphics core with a reduced 128-bit memory interface. To complete the series, the GeForce 9500 GS will launch in July.

 

[Bateluer]

Good to know some exact dates, March 11th.

 

[Extelleron]

If the full 9-series lineup won't even launch until June-July, then it would seem that nVidia's real next-gen card won't launch until sometime late in Q3 or in Q4.

If R700 really launches in June, then nVidia might be in trouble for a few months. Then again, ATI has only been ahead of nVidia on one launch, and that was the X1800->X1900 refresh.

IMO the interesting card there will be the 9800GTX. I think it's been confirmed that it is G92, but still I'd like to see if it will feature a full 384-bit interface w/ 24 ROPS. With a 64SP 9600GT being only 15-30% slower than a 112SP 8800GT, I think it is clear G92 is very limited by its 16 ROPs and 256-bit memory.

 

[nyker96]

I still think the naming scheme for the 9xxx is messed up. It seems that 8800GT is the same exact core and process except with different clocks and 112 stream Proc. this is certainly more powerful than 9600GT and should be named with say a higher number or something. This Nvidia naming scheme is very confusing to us consumers.

 

[Cheex]

I will agree that all eyes will be on the 9800GTX, mine included.

However, if the 9800GT can mimic the price/performance value that the 8800GT brought to the table, then we might very well have another mainstream giant on our hands.

 

[error8]

If the 9800 GTX is a G92 then I don't see it to be much of a performer. No directX 10.1 support and quite probably not much faster then the 8800 GTS 512. In my opinion the 9800 GTX should be at least 2 times faster then the 8800 Ultra, or else I don't see the point of its release. Nvidia should stop using this G92 chip and focus on creating another one. They've released so many cards based on this particular chip, even the 9600 GT is a chopped off 8800 GT, it's becoming strange. If ATI can release a fast HD4000 card, based entirely on a new generation chip, Nvidia can have big problems. It's like AMD when they were in their peak with the Athlon X2 and they just stood there doing nothing until Intel released their core 2 duo chips that destroyed every CPU that AMD had on the market.

 

[Keysplayr]

Hmmm. We are still missing a "G90". We have G92 & G94. G96 will show up soon enough (9200/9300/9500). Wonder what happened to the good 'ol G90. Just thinking out loud here guys.

 

[Cheex]

If G80 is anything to go by, then I'd agree and say the the G90 would be the chip to watch out for.

Question is...will that be the core used in the 9800GTX?

Beats me.

 

[AzN]

Originally posted by: Extelleron
If the full 9-series lineup won't even launch until June-July, then it would seem that nVidia's real next-gen card won't launch until sometime late in Q3 or in Q4.

If R700 really launches in June, then nVidia might be in trouble for a few months. Then again, ATI has only been ahead of nVidia on one launch, and that was the X1800->X1900 refresh.

IMO the interesting card there will be the 9800GTX. I think it's been confirmed that it is G92, but still I'd like to see if it will feature a full 384-bit interface w/ 24 ROPS. With a 64SP 9600GT being only 15-30% slower than a 112SP 8800GT, I think it is clear G92 is very limited by its 16 ROPs and 256-bit memory.

R700 would definitely give G92 chips run for it's money.

G92 definitely needs more memory bandwidth for sure. G92 chip is already big as it is on 65nm process. I'm not so sure about about raising rop count. Let's just hope Nvdia does raise it to 24.

 

[AzN]

Yeah but did Nvidia really have a g82 chip? Maybe they are skipping the whole Gx0 thing.

 

[Miklebud]

Hmmm, my step-up program runs out March 31. I'm crossing my fingers the 9800GTX will be out before then!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

 

[Griswold]

Originally posted by: Extelleron

If R700 really launches in June, then nVidia might be in trouble for a few months. Then again, ATI has only been ahead of nVidia on one launch, and that was the X1800->X1900 refresh.

Did you just forget the 9700pro? Took nvidia almost 6 months (the launch was almost 2 months before actual availablity, IIRC) to release the infamous vacuum cleaner debacle called geforce FX back then, in response to the glorious 9700pro.

Sort of the same we had with G80 and R600, just the other way around. Though, R600 wasnt as bad as the first set of FX cards...

Back to topic!

 

[Modelworks]

Originally posted by: nyker96
I still think the naming scheme for the 9xxx is messed up. It seems that 8800GT is the same exact core and process except with different clocks and 112 stream Proc. this is certainly more powerful than 9600GT and should be named with say a higher number or something. This Nvidia naming scheme is very confusing to us consumers.

Yeah it is confusing.
I wouldn't be surprised to see the 9800xyz and the 9800rsvp one day

 

[nOOky]

What a mess. You have the 8800GT, the 8800GS, and now the 9600GT to really confuse people. Add the 3870 to the mix and wonder why peple get confused.

 

[apoppin]

Originally posted by: Extelleron
If the full 9-series lineup won't even launch until June-July, then it would seem that nVidia's real next-gen card won't launch until sometime late in Q3 or in Q4.

If R700 really launches in June, then nVidia might be in trouble for a few months. Then again, ATI has only been ahead of nVidia on one launch, and that was the X1800->X1900 refresh.

IMO the interesting card there will be the 9800GTX. I think it's been confirmed that it is G92, but still I'd like to see if it will feature a full 384-bit interface w/ 24 ROPS. With a 64SP 9600GT being only 15-30% slower than a 112SP 8800GT, I think it is clear G92 is very limited by its 16 ROPs and 256-bit memory.

What if r700 launches in May?


i appears AMD may be ahead of schedule
:Q

. . . and it happened before x1800>x1900 ... 9700 series dominated

the 9800GTX ultra is the only other card that interests me ... now

 

[lopri]

Found this.

http://www.nordichardware.com/news,7386.html

Judging from the 9600 GT's performance, the G92 core has been tweaked with these new revisions and has a better performance. (9600 GT only has 64 SP) 256-bit memory interface is a bummer, though. 8800 GTS/GT are already using all the bandwidth..

 

[lopri]

If the chart in the link is to be believed, the line-up's look very reminiscent to 7950GX2's time.

9800 GTX SLI > 9800 GX2 > 9800 GTX > 9800 GT
$800 ------------ $500 -------- $ 400 ------ $300

But I gotta say their launch schedule is seriously a mess and probably due changing many times before actual product launch (except the GX2 looks to be set on stone).

Originally posted by: error8
If the 9800 GTX is a G92 then I don't see it to be much of a performer. No directX 10.1 support and quite probably not much faster then the 8800 GTS 512. In my opinion the 9800 GTX should be at least 2 times faster then the 8800 Ultra, or else I don't see the point of its release. Nvidia should stop using this G92 chip and focus on creating another one. They've released so many cards based on this particular chip, even the 9600 GT is a chopped off 8800 GT, it's becoming strange. If ATI can release a fast HD4000 card, based entirely on a new generation chip, Nvidia can have big problems. It's like AMD when they were in their peak with the Athlon X2 and they just stood there doing nothing until Intel released their core 2 duo chips that destroyed every CPU that AMD had on the market.

9800 GTX can be at least 50% faster than 8800 GTX/Ultra. I think the G92 core hasn't shown its true color in current SKUs. As far as DX10.1 is concerned, I suspect that NV might just skip it in an attempt to screw up AMD. Without NV's support, devs will not make games using DX10.1. Microsoft is the only one that can push this thing to both vendors but unfortunately it looks like NV already has been kissing MS' ar**. (Look at APX 2500)

 

[apoppin]

Originally posted by: lopri
Found this.

http://www.nordichardware.com/news,7386.html

Judging from the 9600 GT's performance, the G92 core has been tweaked with these new revisions and has a better performance. (9600 GT only has 64 SP) 256-bit memory interface is a bummer, though. 8800 GTS/GT are already using all the bandwidth..

from a C&P i posted in another thread:

'll let FS conclusion speak to you:

http://www.firingsquad.com/har...performance/page16.asp

Quite honestly, with just 64 stream processors we weren?t expecting much from NVIDIA?s GeForce 9600 GT. We knew it would be a strong competitor to the Radeon HD 3850, but we had no clue it would deliver performance that rivals the GeForce 8800 GT in some cases!

So how did NVIDIA manage to pull it off? Obviously the 9600 GT doesn?t have the pure shading horsepower of the 8800 GT, sporting just 64 stream processors compared to the 8800 GT?s 112, but other than the shading unit deficit the two GPUs are quite similar architecturally. Texture filtering and addressing capabilities are the same, as are the number of ROPs. In addition, they both offer the same peak memory bandwidth and the same z and color compression enhancements. This is important as we?re testing these games with the eye candy cranked up and with AA. In these types of situations, the GPU is often bound by its memory subsystem. Also keep in mind that the 9600 GT sports higher graphics core and stream processor clocks than the 8800 GT.

When you combine this with the OC?ed clocks found on the cards we tested, these 9600 GT cards actually offer more ROP fill and memory bandwidth than a bone stock GeForce 8800 GT board.

well they confirmed my guess
-Not for high resolution with lots of AA/AF

but for $169 a real *deal*

EVGA 9600GT SuperClocked for only $169

 

[lopri]

Is there a review that used a same set of drivers for both 9600 and 8800?

 

[PCTC2]

Originally posted by: lopri

9800 GTX can be at least 50% faster than 8800 GTX/Ultra. I think the G92 core hasn't shown its true color in current SKUs. As far as DX10.1 is concerned, I suspect that NV might just skip it in an attempt to screw up AMD. Without NV's support, devs will not make games using DX10.1. Microsoft is the only one that can push this thing to both vendors but unfortunately it looks like NV already has been kissing MS' ar**. (Look at APX 2500)

Most definitely. Look at the 64SP 256-bit 9600GT competing with the 8800GS and 8800GT 256MB in some areas. If the architecture allows that kind of performance, it can be believed that the G92 has been stubbed so far performance-wise, and with the onset of the 9800GTX, we could expect, if it is a full 384-bit memory system with an optimized G92 core with 128+ SP, a lot more than double the performance of the 9600GT. It could be possible. At $400-500, that would be a performance card I would definitely shell the money out for.

And that would be the reason for the 8800 designations for the current G92 cards, because they offer similar performance shader for shader, but the new 9X00 cards will offer more performance on similar cards but with having less shaders(9600GT vs 8800GT). The 8800GT was simply a snubbed version of the G92 to offer a die-shrink of the older 8800's. The 9X00 series will offer the full performance of the G92 series.... hopefully.

 

[FormulaRedline]

between the end of March and beginning of April

How can you release something between the end of March and the beginning of April? Is there a month I don't know about?

 

[Keysplayr]

Originally posted by: FormulaRedline

between the end of March and beginning of April

How can you release something between the end of March and the beginning of April? Is there a month I don't know about?

Mapril?

 

[the unknown]

Ahhh Mapril.. I've had some good times in Mapril

 

[imported_Shaq]

Originally posted by: FormulaRedline

between the end of March and beginning of April

How can you release something between the end of March and the beginning of April? Is there a month I don't know about?

LOL....between March 11 (GX2) and April 3 (GT) is the real answer, but his sounds much better doesn't it? LOL

 

[apoppin]

Originally posted by: keysplayr2003

Originally posted by: FormulaRedline

between the end of March and beginning of April

How can you release something between the end of March and the beginning of April? Is there a month I don't know about?

Mapril?

i *can* be done ... the instant between 12-midnight March 31 and 12:01:01 April 1

...there is a name for it ... it's called April Fool's release