The official GT300/Fermi Thread

[tdream]

Try a month to a month and a half.

JHH wasn't even holding a real GF100 at the tesla launch. How in the name of god will that timeframe work?

[tajoh111]

So now that the Fermi card is a fake...does this mean that Jen-Hsun Huang will perform the Seppuku?

WTF, this is exactly why AMD fan boys are the absolute worst, they don't know when the line is crossed between being a fan of a company and just being damn hateful.

You want Jen-Hsun Huang to kill himself, then you put a smile face. WTF, thats ing evil.

Anytime a company or person is remotely positive for NV, they are accused for being a fanboy. If this person has any power or any pull, this person is being accused of being paid off by NV. Even Fugger has been accused of being paid off recently.

Man you guys, can't you just stay out of this AMD fans just stop posting in the NV crap in NV threads if your going to just say crap. I have never heard anything positive come from your guys mouths.

Atleast most NV fanboys can say something positive about AMD every once in a while. You guys are just creating a negative atmosphere and making the mods to just give.

The only thing negative about thing I have said about this latest generation is I don't see the 5870 is a worthy upgrade for someone who owns a 4870x2(which is me, and I also have a 4890), and eyefinity is kind of a novelty like 3d gaming because you have to pay alot of money to get it going(more screens), and I am accused of being an NV fanboy when I own a whole lot of AMD stuff.

[Xoulz]

Try a month to a month and a half.

Dude, are you paying attention to whats going on around you..? The GT300 will not be out for the Xmas holiday shopping season..

Nvidia concedes this, just read Anand's review of the Tesla architecture. ATi will have 3 months of un-fettered sales. When the GT300 is released, ATi will simply drop the prices on their current line-up and release (or announce) their X2 line, etc.. competing $ for $.



Do you really think Nvidia will be able to compete with Hemlock (5870x2) @ $599, (come this January or Febuary...) when Nvidia is due to release their GT300.

Or, a 5870 @ $299, or the GT300 for $499..? If Nvidia doesn't have a DX11 $199 part soon, they will loose massive market share and loose $Billions in stock. It doesn't matter how good the GT300 actually is, if only 40k uber-hardcore people ever buy it. The 5870, for the foreseeable future, is good enough for almost anyone needs.

Millions of people will be buying or upgrading their stuff for the Holiday Season. Nvidia will miss out on all those sales.


Timing is everything!

[DilTech]

As promised, sticking to just news posts from here on in this thread.

Not sure if it's been posted guys, but...
http://www.xbitlabs.com/news/video/d...r_Q4_2009.html

It's on track for Q4 '09 according to NVidia.

[Chumbucket843]

JHH wasn't even holding a real GF100 at the tesla launch. How in the name of god will that timeframe work?

thats not an accurate way to speculate when a card will launch.

Even Fugger has been accused of being paid off recently.

its just kind of sad that people would go that far over an email.

Well you know, ATI's the smaller company.

Oh wait, no they're not, because they're part of AMD. NVIDIA's the one fighting the uphill battle here.

indeed, they are going to get screwed over hard in the future if they dont stay ahead. they have to get in embedded markets because intel and amd will have their gpu's on die in pretty much anywhere they can put x86. its not right now that looks iffy for nvidia its in the next few years.im not really worried for them. they have some nice chips from tegra to tesla.

[Particle]

WTF, this is exactly why AMD fan boys are the absolute worst, they don't know when the line is crossed between being a fan of a company and just being damn hateful.

You want Jen-Hsun Huang to kill himself, then you put a smile face. WTF, thats ing evil.

It would be evil if he were serious. He's not literally suggesting the guy should suicide because of GT300 not yet being ready.

No worries. The card will be ready when it's ready. We'll all still be here.

[***Deimos***]

wafer costs are around 3000-5000 us$, so 30 chips per wafer = 100-166$ per gpu, pure die costs

for rv870 it should be around 100 chips per wafer so 30-50$ per chip cost...

these numbers are just examples, they arent accurate...
but you can see, for a rough 50% transistor increase of gt300 over rv870, the costs more or less tripple...

Yields..
vary a lot depending on what libraries are used, logic or SRAM, etc.

Pure specualtion, but we can infer somethings:
- RV670 didn't have a "4830". - this suggests yields were very good.
- RV770 both 4850 and 4870 at launch had 800 SIMD. Few "4830" and company instead opted for seperate 40nm die. Later revision clocks all the way to 1Ghz. - this suggests yields were spectacular
- RV870.. notice for first time in many years the "050" part has cut down shaders etc. Even though die shrink, clockspeed about same. - this suggests yields are poor
Now lets look at nVidia:
- G80. Right away we see GTS variants. Suggests many defects on ROP or MEMCTRL.
- G92. Launced as 112 shader 8800GT. Suggests yields quite poor initially. Couldn't even launch "full" part. But, later we get GTX, and GTX+ and only GTS250 on market. No cut down versions. This indicates yields improved a lot.
- G200. Once again, cut down GTX260 and GTX275 versions. Notice how they "improved" the shaders on GTX260 to 216... yields gradually improving, but because of huge die, many sub-perfect parts.

GT300.
We know 3B transistors. We know 384bit. We know 512 CUDA cores. Thats it. Rumours about yields are just rumours.

But, we do know 3B transistor chip will be bigger than RV870. And AMD launch suggests they are having yield problems. This suggests that *if* and *when* GT300 launches, it might be like G92 launch.

To harvest more chips with defects, 'GTX' could launch with 480 SP, and the GT with 416 SP. Later, after several months, a "Ultra" can be launched with full 512 SP.

Dont be surprised if this happens. They did it with 7800GTX/7800GT. 8800GTX/8800GTS, and ofcourse GTX280/GT260.

[Cybercat]

Or, a 5870 @ $299, or the GT300 for $499..? If Nvidia doesn't have a DX11 $199 part soon, they will loose massive market share and loose $Billions in stock. It doesn't matter how good the GT300 actually is, if only 40k uber-hardcore people ever buy it. The 5870, for the foreseeable future, is good enough for almost anyone needs.

Millions of people will be buying or upgrading their stuff for the Holiday Season. Nvidia will miss out on all those sales.

I think you're exaggerating quite a bit here. Several months without a $200 DX11 card isn't going to cause NVIDIA irreparable damage, or even the millions in lost sales that you claim. The simple truth is people aren't upgrading at the feverish rate that they used to, back when PC gamers had exclusives like HL2, Doom 3, Far Cry and UT2k4 to look forward to. It's not just that the 5870 is "good enough", because not a lot of people will be buying that card either, as it's still in the enthusiast price range. Rather, it's more because there aren't many upcoming bleeding-edge, AAA PC exclusives to galvanize sales.

Besides, this is the first time NVIDIA's been late since 2002. I think they'll survive this generation.

[***Deimos***]

Well, I don't know much about hotlots other than how risky they are. Respins usually take 6-8weeks to get back from TSMC, which means they will get it back the middle of Nov.

So, best case for Nvidia if this next spin is good to go and they start ramping production, they might get a handful of wafers ready for the end of the year, like I said before, maybe a couple hundred cards.


Again... you cannot use an 8+8pin on a "single" GPU, you will not pass PCI SIG certification.

Totally agree about at most a hand full of wafers and couple hundred cards by Dec31/09.

Too many people are idealistic and imagine that because they already have first silicon, its a simple matter to produce boards. While that may be true (making boards is easy), it would not be a sell-able product.

Assuming perfectly working A1 silicon, you dont know its perfectly working until after 3-4 months of extensive testing. Remember, its not as simple as testing a new Coke bottle design, or a desk lamp. If customers indeed will be running scientific applications, all math functionality must be tested precisely. And lets not forget compatiblity with hundreds of monitors, games, engineering and graphic applications. Although some of this can be "dismissed" and fixed in driver later on.

So forgive me if when I see somebody post "board Nov12 - confirmed" that I roll over in laughter.

[EDIT:
To add to what Cybercat said, many people **only** have 20" 1680x1050 displays. Even 8800GT is fine for that with most games. Most people are not crazy obsessive MUST have 8xAA working or x_x... very very few people even have 2560x1600. Last gen (GTX275/4890) already pushes 50+ at 19x10 in virtually every game WITH AA. And its a lot simpler and cheaper to disable a few quality settings than to wait and spend $$$ just to play a $49 game.]

[stangracin3]

- G92. Launced as 112 shader 8800GT. Suggests yields quite poor initially. Couldn't even launch "full" part. But, later we get GTX, and GTX+ and only GTS250 on market. No cut down versions. This indicates yields improved a lot.

8800 GT which got renamed to 9800GT which almost got renamed to a GTS230 or GTS240
then we got 8800GTS which got renamed to 9800GTX which got renamed to 9800GTX+ when overclocked which got renamed to an GTS 250.

[eleeter]

As promised, sticking to just news posts from here on in this thread.

Not sure if it's been posted guys, but...
http://www.xbitlabs.com/news/video/d...r_Q4_2009.html

It's on track for Q4 '09 according to NVidia.

“The first Fermi GPUs are expected to launch by year’s end,” stressed Mr. Alibrandi.

What exactly is meant by "launch"? And then there is the word expected tossed in. Leaves a lot of wiggle room for Nvidia IMO. When Nvidia gives an actual launch/end NDA date, then we can all sit back and expect it. Until then, no one really knows. People are understandably skeptical after the mock up stunt Nvidia pulled.

[villa1n]

“The first Fermi GPUs are expected to launch by year’s end,” stressed Mr. Alibrandi.

What exactly is meant by "launch"? And then there is the word expected tossed in. Leaves a lot of wiggle room for Nvidia IMO. When Nvidia gives an actual launch/end NDA date, then we can all sit back and expect it. Until then, no one really knows. People are understandably skeptical after the mock up stunt Nvidia pulled.

I think people tend to forget, that Nvidia has a share price to Nurse, and to take all "Expected" dates with that in mind... Delays = share price fall, so they as a company have to spin news as much as they can legally, to keep investor confidence up. They may launch some December 31 for posterity, but I think Jan Feb realistically is when we ll see quantity.

[eric66]

probably they meant soft launch i don't expect anything special from nvidia this year

[largon]

highend mainboards can usually supply more than 100W through the slot

Note that slot power and 6/8pin power can never be used to power the same load, thus if a card has an external plug which always provides for GPU VRMs, then slot power cannot be used for powering the same load.

[Jamesrt2004]

As promised, sticking to just news posts from here on in this thread.

Not sure if it's been posted guys, but...
http://www.xbitlabs.com/news/video/d...r_Q4_2009.html

It's on track for Q4 '09 according to NVidia.

didnt THEY say the other day its Q1 and midrange will be "many many months" getting confused now, I reckon late Q4 maybe just before x-mas (if there clever if not just after new year)

[LesGrossman]

3every thread that has to do something with nvidia marketing/physx/sales every ATi fanboy has to get their bashing in and the threads end up being 10+ pages long.

If in trying to defend the defenseless, nvidia fanboys didn't make an absurd amount of absurd posts, the threads would be smaller.

I find it funny how quickly they went from zealots to emmo. "leave nvidia alone", "ATI fans are Bad", "I won't post more" kinda remind me this:
http://www.youtube.com/watch?v=kHmvkRoEowc

[Farinorco]

Oh man, I hate when things turn to a discussion over "why ATi is right and NVIDIA not? You're ATi... You're NVIDIA... you only say that because of hatred... you only say that because of love"... it makes keep the thread so difficult...

I'd like to imagine for a moment that those names belong to a pair of companies that make products for us to buy, like Philips, LG, Woxter, :banana::banana::banana:or, Sony... not to some charismatic heroes of an epic world, or something.

here too is a comparison to crunching on guru.

http://www.guru3d.com/article/radeon...review-test/25

But that comparison is not really a useful one because they are not comparing the same applications / code in different hw, but different code on different hw (note that ATI cards are running code in ATi Stream while NVIDIA cards are running code in CUDA, so that means that the code they are running is different). That's similar to compare the Windows loading time with 2 ssd's (A and B) using Windows XP in ssd A and Windows Vista in ssd B.

1. I love how they leave out RV770, the real competition to GT200, and

2. That is single precision. GT300 will decimate RV870 in double precision.

They are testing HD5870, which is the reviewed card. They are using the most powerful NVIDIA solution, to compare. Which is exactly the point to include the RV770? About the GT300 decimating RV870 in double precision, we will see. That's too much fortune-telling for me. On the other hand I'm not all that interested in the applications that can get a benefit from double precision, since domestic applications and the kind of professional applications that may be useful to me, shouldn't do much about double precision floats.

[Bodkin]

What does ati better/worse have to do with this thread, keep it on topic please

[Helloworld_98]

What does ati better/worse have to do with this thread, keep it on topic please

because ATi has a competing product to GF100.

[Bodkin]

because ATi has a competing product to GF100.

Yeah it does, but gf100 has nothing to do with TWIMTBP when we are looking at it from a technical standpoint. All the bickering should go to the batman thread or one in the GPU section, not a news thread

[Calmatory]

keep it on topic please

No one should post anything but news here then. Let's let this thread sink for the next few weeks and forget about this all.

[BenchZowner]

Why don't we just bash nVIDIA 'til they finally release something ?

[jaredpace]

Anyone see this? I don't know if it's real or not.
http://forums.overclockers.co.uk/sho...&postcount=112

[saaya]

Yields..
vary a lot depending on what libraries are used, logic or SRAM, etc.

Pure specualtion, but we can infer somethings:
- RV670 didn't have a "4830". - this suggests yields were very good.
- RV770 both 4850 and 4870 at launch had 800 SIMD. Few "4830" and company instead opted for seperate 40nm die. Later revision clocks all the way to 1Ghz. - this suggests yields were spectacular
- RV870.. notice for first time in many years the "050" part has cut down shaders etc. Even though die shrink, clockspeed about same. - this suggests yields are poor
Now lets look at nVidia:
- G80. Right away we see GTS variants. Suggests many defects on ROP or MEMCTRL.
- G92. Launced as 112 shader 8800GT. Suggests yields quite poor initially. Couldn't even launch "full" part. But, later we get GTX, and GTX+ and only GTS250 on market. No cut down versions. This indicates yields improved a lot.
- G200. Once again, cut down GTX260 and GTX275 versions. Notice how they "improved" the shaders on GTX260 to 216... yields gradually improving, but because of huge die, many sub-perfect parts.

GT300.
We know 3B transistors. We know 384bit. We know 512 CUDA cores. Thats it. Rumours about yields are just rumours.

But, we do know 3B transistor chip will be bigger than RV870. And AMD launch suggests they are having yield problems. This suggests that *if* and *when* GT300 launches, it might be like G92 launch.

To harvest more chips with defects, 'GTX' could launch with 480 SP, and the GT with 416 SP. Later, after several months, a "Ultra" can be launched with full 512 SP.

Dont be surprised if this happens. They did it with 7800GTX/7800GT. 8800GTX/8800GTS, and ofcourse GTX280/GT260.

very nice post, thx, and also thx for bringing this thread back on topic!

and yes, i agree, i wouldnt be surprised if nvidia launches a cut down gpu card first. all they need right now is to beat a 5870 and a possible overclocked rv870 aka 5890... if they dont need all sps for that, they probably wont enable all sps initially.

Note that slot power and 6/8pin power can never be used to power the same load, thus if a card has an external plug which always provides for GPU VRMs, then slot power cannot be used for powering the same load.

oh really? hmmm thx i didnt know that...
im confused though, what do cards use the slot power for then?
memory pwm? but 75W for memory, isnt that way too much?

Anyone see this? I don't know if it's real or not.
http://forums.overclockers.co.uk/sho...&postcount=112

[ajaidev]

Anyone see this? I don't know if it's real or not.
http://forums.overclockers.co.uk/sho...&postcount=112

Yep, looks real

But a 20 pin is rather too old, they should have use the 24 pin connector instead The USB ports are a added plus add a gamepad on ur GPU it self lesser latency related lag