Sunday, September 17th 2017

NVIDIA Segregates Turing GPUs; Factory Overclocking Forbidden on the Cheaper Variant

While working on GPU-Z support for NVIDIA's RTX 20-series graphics cards, we noticed something curious. Each GPU model has not one, but two device IDs assigned to it. A device ID is a unique identification that tells Windows which specific device is installed, so it can select and load the relevant driver software. It also tells the driver, which commands to send to the chip, as they vary between generations. Last but not least, the device ID can be used to enable or lock certain features, for example in the professional space. Two device IDs per GPU is very unusual. For example, all GTX 1080 Ti cards, whether reference or custom design, are marked as 1B06. Titan Xp on the other hand, which uses the same physical GPU, is marked as 1B02. NVIDIA has always used just one ID per SKU, no matter if custom-design, reference or Founders Edition.

We reached out to industry sources and confirmed that for Turing, NVIDIA is creating two device IDs per GPU to correspond to two different ASIC codes per GPU model (for example, TU102-300 and TU102-300-A for the RTX 2080 Ti). The Turing -300 variant is designated to be used on cards targeting the MSRP price point, while the 300-A variant is for use on custom-design, overclocked cards. Both are the same physical chip, just separated by binning, and pricing, which means NVIDIA pretests all GPUs and sorts them by properties such as overclocking potential, power efficiency, etc.
When a board partner uses a -300 Turing GPU variant, factory overclocking is forbidden. Only the more expensive -30-A variants are meant for this scenario. Both can be overclocked manually though, by the user, but it's likely that the overclocking potential on the lower bin won't be as high as on the higher rated chips. Separate device IDs could also prevent consumers from buying the cheapest card, with reference clocks, and flashing it with the BIOS from a faster factory-overclocked variant of that card (think buying an MSI Gaming card and flashing it with the BIOS of Gaming X).

All Founders Edition and custom designs that we could look at so far, use the same -300-A GPU variant, which means the device ID is not used to separate Founders Edition from custom design cards.
Add your own comment

81 Comments on NVIDIA Segregates Turing GPUs; Factory Overclocking Forbidden on the Cheaper Variant

#1
londiste
I have only one question: Where the hell are all these -400 Turing GPU's at MSRP then, huh? :D
Posted on Reply
#2
FreedomEclipse
~Technological Technocrat~
So....

In short, we are going to be needing software mods to get overclocking back?? Challenge accepted

(not by me of course, I wouldnt know where to start but there are lots of smart people on the internet)



Ahhh I get it now.
Posted on Reply
#3
Vya Domus
To me this is just another sign that Nvidia are slowly trying to move onto manufacturing their own cards independently.
Posted on Reply
#4
btarunr
Editor & Senior Moderator
FreedomEclipse said:
So....

In short, we are going to be needing software mods to get overclocking back??
Manual overclocking by end-users (you) allowed on all cards.
Posted on Reply
#5
jabbadap
FreedomEclipse said:
So....

In short, we are going to be needing software mods to get overclocking back?? Challenge accepted

(not by me of course, I wouldnt know where to start but there are lots of smart people on the internet)
Read again, they are like gtx 1070ti, AIBs are prohibited to sell OC variants of lesser binned chip. There's no restrictions to end users.
Posted on Reply
#6
the54thvoid
If they're restricting overclocking on cards that cannot handle it, is that so bad? There's a lot of silicon in them there hills, may be it's temperamental?
What sucks is the premium for the ones that can overclock. I'll be surprised if the cheap one is cheaper than the MSRP.
Posted on Reply
#7
Solidstate89
This seems like a good thing that they're basically doing their own in-house binning for every chip that gets released as opposed to relying on the AIB manufacturers to bin the chips for their high-end parts only. Manual OC'ing is still allowed for everything but factory overclocks are exclusive to the binned chips which probably also means they'll manually OC better for the user as well.
Posted on Reply
#8
Aquinus
Resident Wat-man
Vya Domus said:
To me this is just another sign that Nvidia are slowly trying to move onto manufacturing their own cards independently.
That is a monumental feat that will likely backfire in nVidia's face if they try to do this. Owning your own factories to produce chips like these is an industry in and of itself. If Intel seems to be struggling to do it, what makes you think nVidia is going to fare any better? It's not like Intel has no experience doing this either so, it just doesn't feel like the right move which makes me question the validity of your statement.
the54thvoid said:
If they're restricting overclocking on cards that cannot handle it, is that so bad? There's a lot of silicon in them there hills, may be it's temperamental?
Maybe a reason why reviews keep getting pushed out? Driving enhancements to make the GPUs not look as bad as they really are. I think they're buying time to save face for rushing a product out the door and over-hyping it. The reality is that there was no need for it. I honestly think nVidia has more to lose than not by pulling this kind of stunt.
Posted on Reply
#9
Caring1
What's going to stop the Device I.D. being changed?
It's been done before to fake cards being something they aren't.
Posted on Reply
#10
dj-electric
I'm afraid this has something to do with their AI overclocking.
Gonna test it with my third-party MSI card and see what happens.
Posted on Reply
#11
ZoneDymo
the54thvoid said:
If they're restricting overclocking on cards that cannot handle it, is that so bad? There's a lot of silicon in them there hills, may be it's temperamental?
What sucks is the premium for the ones that can overclock. I'll be surprised if the cheap one is cheaper than the MSRP.
Well.... its them (possibly) claiming they cannot handle the overclock.
Don't you think a manufacture would find out by themselves if it can or cannot handle an overclock?

If Nvidia said "our lower-end cards wont be able to be overclocked, they will become unstable", that would be an entirely different story then "we FORBID YOU to sell overclocked cards of this variant".
Posted on Reply
#12
kings
Aquinus said:
The reality is that there was no need for it.
Well, It was people who badly wanted a new series of graphics cards in 2018. They already have prepared the pitchforks...

The fact is, Nvidia did not have to launch anything. Pascal, despite the age, would continue being the best high-end cards on the market.
Posted on Reply
#13
Vya Domus
Aquinus said:
That is a monumental feat that will likely backfire in nVidia's face if they try to do this. Owning your own factories to produce chips like these is an industry in and of itself. If Intel seems to be struggling to do it, what makes you think nVidia is going to fare any better? It's not like Intel has no experience doing this either so, it just doesn't feel like the right move which makes me question the validity of your statement.
I am not talking about manufacturing their own silicon I am talking about the cards themselves. Nvidia keeps pushing this Founders Edition thing and it looks to be in the detriment of custom designs from AIBs. If you think about it there is no reason any custom card would be superior in any significant way over the FE variants other than the price and even then it turns out they're all going to be massivly overpriced.

Why the hell would you pay 1200$ for a blower 2080ti (with potentially restricted overclocking) and not buy it straight from Nvidia for roughly the same money and with a very good cooler, power delivery and all that. See where this is going ? Of course right now they have nowhere near the required capacity to produce enough cards but they are sure taking measures to ensure that when the time comes they'll have a easy time. It all makes a lot of sense when you look into it, that's the next big step Nvidia wants to make.
Posted on Reply
#14
W1zzard
Caring1 said:
What's going to stop the Device I.D. being changed?
It's been done before to fake cards being something they aren't.
This was fixed many years ago. The fakes are based on GTS 450 and GTX 550 which could have their IDs changd to anything else
Posted on Reply
#15
iO
So they found an additional income by charging a premium for binned chips if the AIBs want to offer OC models.
And we customer are no longer able to get a cheap card and OC it for extra performance. Yay!
Posted on Reply
#16
TheinsanegamerN
the54thvoid said:
If they're restricting overclocking on cards that cannot handle it, is that so bad? There's a lot of silicon in them there hills, may be it's temperamental?
What sucks is the premium for the ones that can overclock. I'll be surprised if the cheap one is cheaper than the MSRP.
And what guarantee is there that Nvidia is only binning chips that cant OC, as opposed to, you know, only allowing OCs on these Special chips that cost more.

This is nvidia we are talking about here, if they can fleece people for more money, they will do it.
Posted on Reply
#17
Gasaraki
OK... Thanks for the clickbait headline.

Nothing to see here folks, move along.
Posted on Reply
#18
chaosmassive
so it'll be RTX 2080 Ti
and...RTX 2080 Ti-K UNLOCKED, UNLEASHED, UNRESTRAINED !! version
nvidia eyes greener than ever !
Posted on Reply
#19
Solidstate89
Gasaraki said:
OK... Thanks for the clickbait headline.

Nothing to see here folks, move along.
How is it clickbait? The headline describes exactly what is happening. Their AIB partners aren't allowed to sell factory overclocked cards unless they use the specifically binned chips.
Posted on Reply
#20
rtwjunkie
PC Gaming Enthusiast
Vya Domus said:
I am not talking about manufacturing their own silicon I am talking about the cards themselves. Nvidia keeps pushing this Founders Edition thing and it looks to be in the detriment of custom designs from AIBs
What blower cooler? Founders Editions are not blower coolers with 20 series. Did you miss all the pictures?
Posted on Reply
#21
medi01
I think market does indeed deserve "nvidia's K chips".

My only problem is with how "serial" will the stuff that reviewers, who signed draconian NDA, will publish on 19th be.

FreedomEclipse said:
So....

In short, we are going to be needing software mods to get overclocking back?? Challenge accepted

(not by me of course, I wouldnt know where to start but there are lots of smart people on the internet)


Ahhh I get it now.
Erm, there were plenty of female nazis, the most famous director, for once, what is your signature about?
Posted on Reply
#22
coonbro
for the extremely low pricing of these cards to start with I can see this as a way to increase the profit platform . in a way you kinda seen this with say evga and a ref. sc may oc well or not at all , but a classy / kingpin is binned to hold there best factory oc + what you can add . maybe I'm wrong but it sounds like what evga has been doing in there lines of cards anyway now just on nvidias own guideline scale
Posted on Reply
#23
TheinsanegamerN
rtwjunkie said:
What blower cooler? Founders Editions are not blower coolers with 20 series. Did you miss all the pictures?
I think he is referring to AIBs that makes plastic blowers of reference cards that are cheaper then FEs, and why would you buy an AIB blower when the reference nvidia cooler is as good as third party coolers. In such a situation, there would be no real reason to buy a third party nvidia card at all, as the reference card is already as good as AIB cards. It comes off as nvidia gunning to be like 3DFX, making all their own cards.
Posted on Reply
#24
xorbe
Are they actually binned, or does the -A simply denote that the oc tax was paid?
Posted on Reply
#25
B-Real
btarunr said:
Manual overclocking by end-users (you) allowed on all cards.
And I bet that will influence the warranty and they will say "you OCd it, no rebate/repair".
Posted on Reply
Add your own comment