NVIDIA GK104 Coming Soon?

Status
Not open for further replies.
"should be faster?" "in most cases"? damn right it should be faster since always second to come has time to push few Mhz to gain the extra 4,5 fps :)) i like how that post tries to convince u of something,but is not really convinced himself.Big news anyway,like nvidia would release few months later a slower card....
 
According to THIS rumour new GTX680 will be GK104, 1536 SP, etc. and will be shown 12th March, released 23rd March.
I thought GK110 was going to be GTX680...

What I've read elsewhere; to be shown to a few select journalists on 12th and NDA ends 23rd.
 
http://www.4gamer.net/games/120/G012093/20120306077/

That article there (in Japanese) may explain how GK104 will manage to surpass both the gtx580 and hd7970 in performance, while being smaller than Tahiti. Hint: gpu-based "hyperthreading" It's difficult to tell exactly what the article is saying due to the translation, but seems to be full of some good information nonetheless.
 
Strange indeed. Its memory bandwidth points to it being a mid range card, almost the same as GTX 580

You'd expect the 680 to have more bandwidth.

I thought the GK104 was the mid-range card, behind the eventual GK110.
 
http://www.4gamer.net/games/120/G012093/20120306077/

That article there (in Japanese) may explain how GK104 will manage to surpass both the gtx580 and hd7970 in performance, while being smaller than Tahiti. Hint: gpu-based "hyperthreading" It's difficult to tell exactly what the article is saying due to the translation, but seems to be full of some good information nonetheless.

Translation from Google:

"I think the details are want to tell once again the opportunity, as can be seen if Moraere Think of a Hyper-Threading "" of Intel "approach is NVIDIA, to be able to use threading to other free resources CUDA Core It is, and shall raise the processing performance per CUDA Core, be reduced so that the power consumption per performance "that (official NVIDIA).
In addition, officials close to NVIDIA, "in Kepler, and to improve the processing performance of the DirectCompute PhysX" but also said, If we take the approach NVIDIA officials say certainly, than the arithmetic processing graphics processing "light" is not assigned to the free resources of CUDA Core, a, would not be surprised as to improve the performance be measured...."
 
I thought the GK104 was the mid-range card, behind the eventual GK110.

It is the mid-range card/chip, despite how people are trying to spin the facts in this thread.

"should be faster?" "in most cases"? damn right it should be faster since always second to come has time to push few Mhz to gain the extra 4,5 fps :)) i like how that post tries to convince u of something,but is not really convinced himself.Big news anyway,like nvidia would release few months later a slower card....


That nVidia's mid-range is going to compete with/beat AMD's top-end is pretty impressive, especially only a couple of months later.
 
Last edited:
What I've read elsewhere; to be shown to a few select journalists on 12th and NDA ends 23rd.

The fud article calls it a GTX 680 press event, which hopefully would mean a paper launch and not just an NDA showing. Makes sense for Nvidia to announce something before the 78xx cards launch (if the Mar 23 retail date is right for Kepler), especially if they are planning on pricing it aggresively.

Wasn't DX11 performance already one of the 79xx cards' strong points? If Kepler really can beat them in DX11, that would be impressive.
 
I just want the cards to come out this month or first week of April. End of story.

Honestly, I can't remember any other topic that has had more speculation than graphics cards in the past several years of technology... maybe Apple products? But who cares about those.
 
maybe Apple products? But who cares about those.

Apparently a metric shit-ton, based on their sales numbers.

Let's face it, the GPU is about the most exciting part of the PC that gets updated frequently. No one cares about next-gen hard drives or monitors, and even CPUs don't have the same draw (maybe since performance increases are so steady - you don't get the wow! moment like you sometimes do with GPUs).
 
Apparently a metric shit-ton, based on their sales numbers.

Let's face it, the GPU is about the most exciting part of the PC that gets updated frequently. No one cares about next-gen hard drives or monitors, and even CPUs don't have the same draw (maybe since performance increases are so steady - you don't get the wow! moment like you sometimes do with GPUs).

I blame that on blind following.

But yeah, GPUs are definitely the most talked about computer part. The only things behind it I can think of is the computer case and the CPU.
 
Let's face it, the GPU is about the most exciting part of the PC that gets updated frequently. No one cares about next-gen hard drives or monitors, and even CPUs don't have the same draw (maybe since performance increases are so steady - you don't get the wow! moment like you sometimes do with GPUs).

Well, the reason for that is that I would say the vast majority of enthusiasts build high-end machines to either 1) play games or 2) crunch numbers, or both, and the GPU being the primary performance/feature indicator of those two tasks ends up being the most "important" or "exciting" part.

New hard drive technology is probably exciting to enterprise server admins. :p

CPUs actually used to be more exciting before they started using GPGPUs to crunch numbers in supercomputers.
 

It's FUD-tacular tripe:

Nvidia on the other hand is announcing their newest family of parts next Monday. It is a good part, but supply is the proverbial elephant in the room. Ivy Bridge is launching soon, and if GPUs aren’t ready, things get ugly in a hurry. Given Nvidia’s uncharacteristic whining during the last conference call, you now understand what they were prepping the analyst community for. SemiAccurate still believes that Nvidia is the only company seeing 28nm ‘yield problems’ and the real problem has it’s roots in Santa Clara.

All in all, this is nothing more than a horrific failure. Complete 28nm stoppage at TSMC with no public word, no explanation, and no known concrete resumption date. Sources tell us that it is indeed temporary, but until there is a good explanation, don’t assume anything.S|A

In one sentence Charlie trolls about nVidia being alone in their 28nm stoppage, then the next he boasts about how it's for all of TSMC and not just nVidia... just tabloid material.
 
Well that isn't good. That could be a major delay, but I guess it's good that they're fixing it rather than releasing another Fermi.

TSMC wouldn't stop production if the problem was limited just to Nvidia chips. Therefore, it's not Nvidia's problem to fix and we can all assume GK104 isn't GF100 all over again. There aren't reports of Kepler only yielding 2%, being too big, too hot, or too power hungry.
 
It's FUD-tacular tripe:

In one sentence Charlie trolls about nVidia being alone in their 28nm stoppage, then the next he boasts about how it's for all of TSMC and not just nVidia... just tabloid material.

And what does Ivy launching have to do with Kepler and things getting ugly? The two are completely unrelated.
 
And what does Ivy launching have to do with Kepler and things getting ugly? The two are completely unrelated.

because the moment ivy chips are in consumers' hands, they will want to upgrade.

if kepler isnt in their hands at that same time, regardless of whether it beats 7970 or 7990, it becomes a gigantic launch failure for nvidia, because those upgrading to ivy will likely be looking at upgrading to the new gen of video cards. If Kep isnt ready when Ivy is, nvidia misses out on tons of money.

it was a smart observation on his part.
 
because the moment ivy chips are in consumers' hands, they will want to upgrade.

if kepler isnt in their hands at that same time, regardless of whether it beats 7970 or 7990, it becomes a gigantic launch failure for nvidia, because those upgrading to ivy will likely be looking at upgrading to the new gen of video cards. If Kep isnt ready when Ivy is, nvidia misses out on tons of money.

it was a smart observation on his part.

New cards weren't launched when SB launched, and that wasn't a huge failure for anybody, plus IB is a smaller upgrade than SB was. It may have some impact on mobile parts, since laptop makers may want to include both tech in new products, but there appear to already be some mobile 28nm parts already shipping, so that seems like a non-factor. I don't believe that there is a huge cadre of people waiting to upgrade both th CPU and GPU when Ivy launches - enthusiasts who are spending that kind of money have no problem de-linking those purchases.
 
New cards weren't launched when SB launched, and that wasn't a huge failure for anybody, plus IB is a smaller upgrade than SB was. It may have some impact on mobile parts, since laptop makers may want to include both tech in new products, but there appear to already be some mobile 28nm parts already shipping, so that seems like a non-factor. I don't believe that there is a huge cadre of people waiting to upgrade both th CPU and GPU when Ivy launches - enthusiasts who are spending that kind of money have no problem de-linking those purchases.

you dont get it
there are people who dont have sandy who are waiting to upgrade their system to ivy

they are waiting for ivy launch.

if ivy launches, and they still have to choose between 7870--7970 or a paper launch of kep, where do you think they will go?
 
you dont get it
there are people who dont have sandy who are waiting to upgrade their system to ivy

they are waiting for ivy launch.

if ivy launches, and they still have to choose between 7870--7970 or a paper launch of kep, where do you think they will go?

Maybe they'll upgrade the CPU and wait another month or so for the video card? Why would you think they'd have to do both at the same time? Why is it not big problems for Intel that IB wasn't ready when Tahiti launched? Plenty of people were waiting to upgrade to that also.

I just don't see that Kepler sales are tied to Ivy Bridge sales in any meaningful way. People who are savvy enough to be waiting for Ivy Bridge are probably also savvy enough to know what is happening in the video card realm and be willing to wait for what they want.
 
Maybe they'll upgrade the CPU and wait another month or so for the video card? Why would you think they'd have to do both at the same time? Why is it not big problems for Intel that IB wasn't ready when Tahiti launched? Plenty of people were waiting to upgrade to that also.

I just don't see that Kepler sales are tied to Ivy Bridge sales in any meaningful way. People who are savvy enough to be waiting for Ivy Bridge are probably also savvy enough to know what is happening in the video card realm and be willing to wait for what they want.

You would be surprised. There is a breaking point and you can't wait forever. People who are thinking of building a new system generally do it when new CPU becomes available. They also tend to buy whatever is available, because they want everything there ready when they are doing the build.

As for your sandybridge reference, well both Nvidia and AMD had just released their video cards the month before Sandybridge. All I know is sales of video cards increase for me around the release of a new CPU.
 
You would be surprised. There is a breaking point and you can't wait forever. People who are thinking of building a new system generally do it when new CPU becomes available. They also tend to buy whatever is available, because they want everything there ready when they are doing the build.

As for your sandybridge reference, well both Nvidia and AMD had just released their video cards the month before Sandybridge. All I know is sales of video cards increase for me around the release of a new CPU.

I certainly agree that CPUs tend to drive the upgrade cycle, and I don't have any empirical data either way, so if you do I'll take your word for it, but I certainly don't get the sense that there are a lot of people waiting for Ivy with the expectation that they would force themselves into a video card purchase at the same time. I can certainly believe that a lot of people doing a CPU/MB upgrade might also do a GPU upgrade at the same time, I'm just not convinced that it would cause significant sales losses for Nvidia (especially since supply will be tight no matter what and HD 79xxs aren't exactly stacked to the ceilings).

I think as long as Nividia paper-launches the card this month, they'll be able to freeze enough buyers to be happy.
 
I'm with Forceman, I suspect the correlation is overblown here.
 
New CPUs mean big wave of new OEM designs. Its in ATI/Nvidia's best interest to have GPUs ready. Probably why we heard of GK107 way before GK104, laptop designs.

General Manager of mobile division at Nvidia Rene Haas, told Fudzilla that Kepler has more design wins than Fermi. He was talking notebooks of course as this is what he does.

He didn’t go into details on when Kepler launches but most of our meetings at CES indicated that mobile Kepler should come around Ivy Bridge launch time.
 
New CPUs mean big wave of new OEM designs. Its in ATI/Nvidia's best interest to have GPUs ready. Probably why we heard of GK107 way before GK104, laptop designs.

Now that's a meaningful observation. Profit margin on consumer parts is fat, but volume is all in the OEM market and once you get a design win, it's a nice steady revenue stream.
 
Maybe they'll upgrade the CPU and wait another month or so for the video card? Why would you think they'd have to do both at the same time? Why is it not big problems for Intel that IB wasn't ready when Tahiti launched? Plenty of people were waiting to upgrade to that also.

I just don't see that Kepler sales are tied to Ivy Bridge sales in any meaningful way. People who are savvy enough to be waiting for Ivy Bridge are probably also savvy enough to know what is happening in the video card realm and be willing to wait for what they want.

Some people will, like myself (see sig) but in general, if they have the money to spend, and there's the AMD card and nVidia is still "vaporware", well, you get the picture...
 
New CPUs mean big wave of new OEM designs. Its in ATI/Nvidia's best interest to have GPUs ready. Probably why we heard of GK107 way before GK104, laptop designs.

^^ Another good observation. We nut ball upgraders are a minority...oem's are where the cash is at.
 
It's all about the money, if nivida cant release cards with 10% better performance for equalish price that's whats gonna happen. AMD really screw consumers with the prices of these cards. Years ago they made nvidia do massive price drops on $600 cards they just released. Now there the ones released expensive cards.
 
in the site there is an article that says that samaritan demo runned on a single kepler instead of a sli of 580, this can let us think that kepler is twice fast than a 580.
 
in the site there is an article that says that samaritan demo runned on a single kepler instead of a sli of 580, this can let us think that kepler is twice fast than a 580.

What they actually said was it was running on 3 x 580 before, and now one Kepler. I don't think anyone actually believes it's 3x faster than a 580, though.
 
all that article is going to do is stir up crap and confuse people. the first time they ran the demo was using 2 gtx580 cards for graphics and 1 gtx580 for physx. when they ran it this week, it was on 1 gtx680 but they also only used FXAA instead of MSAA.
 
Popped up on google search yesterday.

8 days and 10 hours, ill will spill some serious beans...


all i can say is NVIDA surround will not run on one card, heaven benchmark with everything maxed above 29fps in surround, thats all i can say, believe or not


wow i don't know why my phone auto corrected to NOT, its obviously wrong

Last 24 hours.

look what i have been playing with for the last 5 weeks

2012-03-15_135626-1.jpg



These things are out and about in key players hands.
 
Card looks nice...but im gonna have to do something about that geforce logo. too big and too green. lol
 
Status
Not open for further replies.
Back
Top