Friday, September 11th 2020

NVIDIA GeForce RTX Ampere Chips Feature Three Binning Tiers, Mostly Good Dies are Present

Chip binning is a process of sorting out the manufactured silicon by quality. That means that each chip that comes from the silicon wafer is tested and sorted by different features. For example, a chip is tested for how much voltage it takes for operation, how cool it runs, and of course how it overclocks. By putting their chips through various testing, manufacturers often create binning tiers, where they can differentiate good and bad chips, so they know where to send, and if they should send the chips. The biggest and most complex approach for sending chips is for graphics cards. As there are different AIBs, manufacturers like NVIDIA and AMD need to send them chips of various qualities to incorporate in their products. It is a rather time-consuming and complex process to find out the bin type and the tier of chips, however today we are getting some information from Igor's Lab.

According to their sources, it is said that NVIDIA's latest GeForce RTX Ampere lineup features three binning tiers. There is "Bin 0" which represents an okay chip that can perform as intended, "Bin 1" chips which are good processors, and "Bin 2" processors which represent the best quality chips with the highest performance characteristics. These "Bin 2" dies run cooler compared to the rest and achieve higher overclocking speed. In reality, the binning represents coordination between the chip designer (NVIDIA in this case) and the manufacturer (Samsung with its 8N 8 nm process). It is said that from the complete pilot run of Ampere chips, Samsung ends up with 30% of the "Bin 0" dies, 60% of "Bin 1" dies, and only 10% of "Bin 2" dies. The production period was quite short and these numbers are good for Samsung, as they probably didn't have much time to work on it, so we can expect these numbers to improve.
Source: Igor's Lab
Add your own comment

66 Comments on NVIDIA GeForce RTX Ampere Chips Feature Three Binning Tiers, Mostly Good Dies are Present

#51
TheoneandonlyMrK
Vayra86And yet the same guy says right below you that most of the time you really don't know. Even if there is a difference, the gap is minute at best and extreme OC crowds are not what the majority of these people are posting as. They are just posting as consumers, and those I do care about, because this is once again a whole lot of ado about nothing people get confused about. Bench score ninjas are what, 1%?

'Is my RTX a worse one' omg must buy top end version. That is the sentiment that is being falsely created here.

You'd also have to wonder when that point arrives where 'later in the lifecycle' it suddenly won't matter. Lottery is lottery. End of story. And all those stories about 'omg this stepping and this period' are very old, well known, and we've arrived at a point in foundry refinement that these things are nowhere near what they used to be. If the total performance gap is 2-3% you're already being very generous. I'm a big promotor of such perspective when these claims are made in news. Today the vast majority of this hype is just that, lots of hobbyists riding the train together.

But the most telling of it all, is that so far.... no sources were provided.
I rest my case and this article identifies as clickbait.
While performance differential could be argued Nvidia definitely do bin chip's, that's a known fact?!
And since they actually label them differently, is it arguable ?, we had this with Turing so it's happening.
New news it's not.
Posted on Reply
#52
Dyatlov A
How to know i am buying the best bin 2? I will need to use with 450W PSU, therefore very important to me good undervolting capabilities.
Posted on Reply
#53
Assimilator
Dyatlov AHow to know i am buying the best bin 2? I will need to use with 450W PSU, therefore very important to me good undervolting capabilities.
Buy a better PSU or don't buy an RTX 3000 card.
Posted on Reply
#54
Vayra86
theoneandonlymrkWhile performance differential could be argued Nvidia definitely do bin chip's, that's a known fact?!
And since they actually label them differently, is it arguable ?, we had this with Turing so it's happening.
New news it's not.
They labelled a few chips with A- versions indeed.

Is that the case here? I didn't read it. In fact I didn't read anything much of substance here. Devil as always is in the details...

In the same vein, we 'had this with Turing' but all we really had was a set of chips that got blessed with higher TDP budgets alongside some vagueness about binning.
Hey look, if I tweak the power target with any Nvidia card with GPU Boost, I get different numbers too. We all know the boost algorithm finds space within the aforementioned TDP budget. A card sold with higher TDP budget simply has more headroom to boost, and that is what they managed to put on a box too. Wow. Such Binning. Much performance. :roll::roll::roll:

People oughta wake the f up...
Posted on Reply
#55
TheoneandonlyMrK
Vayra86They labelled a few chips with A- versions indeed.

Is that the case here? I didn't read it. In fact I didn't read anything much of substance here. Devil as always is in the details...

In the same vein, we 'had this with Turing' but all we really had was a set of chips that got blessed with higher TDP budgets alongside some vagueness about binning.
Hey look, if I tweak the power target with any Nvidia card with GPU Boost, I get different numbers too. We all know the boost algorithm finds space within the aforementioned TDP budget. A card sold with higher TDP budget simply has more headroom to boost, and that is what they managed to put on a box too. Wow. Such Binning. Much performance. :roll::roll::roll:

People oughta wake the f up...
So it's all bullshit then?

They don't find the less leaky chips and pass those to the quadro team/buyer's?.

They then just have big bins of working and not working and it's lucky dip?.

Riiight, we disagree but do note I wasn't asleep , give yourself a nudge though, just check.
Posted on Reply
#56
Assimilator
theoneandonlymrkSo it's all bullshit then?

They don't find the less leaky chips and pass those to the quadro team/buyer's?.

They then just have big bins of working and not working and it's lucky dip?.

Riiight, we disagree but do note I wasn't asleep , give yourself a nudge though, just check.
What exactly does Quadro have to do with any of this?

And of course there's binning. It's just that it won't matter to most consumers. Unless they want to pay more for a better-binned card. Which NVIDIA is perfectly within its rights to charge them extra for.

What will happen is that, since NVIDIA won't be able to sell 10% of its best-binned chips at a higher price (maybe 1%), most of them will end up going into the ordinary consumer pool, and the silicon lottery will work its democratic nature as always. In other words, it'll be just like Turing, and Pascal, and every series before...
Posted on Reply
#57
TheoneandonlyMrK
AssimilatorWhat exactly does Quadro have to do with any of this?

And of course there's binning. It's just that it won't matter to most consumers. Unless they want to pay more for a better-binned card. Which NVIDIA is perfectly within its rights to charge them extra for.

What will happen is that, since NVIDIA won't be able to sell 10% of its best-binned chips at a higher price (maybe 1%), most of them will end up going into the ordinary consumer pool, and the silicon lottery will work its democratic nature as always. In other words, it'll be just like Turing, and Pascal, and every series before...
Yes where the best chip's , with the lowest leakage and operating volts go to enterprise/Quadro see the link.
Posted on Reply
#58
Assimilator
theoneandonlymrkYes where the best chip's , with the lowest leakage and operating volts go to enterprise/Quadro see the link.
What link? The OP says nothing about Quadro.
Posted on Reply
#59
TheoneandonlyMrK
AssimilatorWhat link? The OP says nothing about Quadro.
There's Links between Quadro and binning chips not a link to the Op, I originally replied to Vayra not you.
Posted on Reply
#60
Vayra86
theoneandonlymrkSo it's all bullshit then?

They don't find the less leaky chips and pass those to the quadro team/buyer's?.

They then just have big bins of working and not working and it's lucky dip?.

Riiight, we disagree but do note I wasn't asleep , give yourself a nudge though, just check.
Nobody is denying binning... I am denying the need for this whole article to exist. What's being reported here? What's the point? This is like saying 'Cars run on petrol... but some run on electricity or gas'. Well, thanks for the update :D

Water is wet but here we are being led to believe something special is going on. OMG they have three tiers 'what if I get a bad one', 'how do I find a good one'...

Fake demand & clickbait. Look at the headline, 2nd part. 'Mostly good dies are present'... are we insinuating you might strike on a bad one with your purchase?! Nope... But its still in the title. Does this then mean most wafers are nearly perfect with just 'somewhat less perfect' dies on them? Obviously not. So what is it then?

So yes its all bullshit and many fools and money are involved. Any chip that is binned for a specific product is binned for that product, its the most basic thing since forever. I don't need to click links for that. Any time Nvidia thinks a specific die may go into another product, they actually tell you. We've had our share of GP104 movements for example with Pascal. We had A-dies with Turing. In neither case though did those qualify for a product that performed differently from the similar named one with a different chip inside. Even the so called remarkable Turing FE's needed a higher power target to 'be better'. Nuff said.
Posted on Reply
#61
londiste
theoneandonlymrkYes where the best chip's , with the lowest leakage and operating volts go to enterprise/Quadro see the link.
Lowest leakage and operating volts may not - and often do not - coincide with best overclockability.
Posted on Reply
#62
TheoneandonlyMrK
Vayra86Nobody is denying binning... I am denying the need for this whole article to exist. What's being reported here? What's the point? This is like saying 'Cars run on petrol... but some run on electricity or gas'. Well, thanks for the update :D

Water is wet but here we are being led to believe something special is going on. OMG they have three tiers 'what if I get a bad one', 'how do I find a good one'...

Fake demand & clickbait. Look at the headline, 2nd part. 'Mostly good dies are present'... are we insinuating you might strike on a bad one with your purchase?! Nope... But its still in the title. Does this then mean most wafers are nearly perfect with just 'somewhat less perfect' dies on them? Obviously not. So what is it then?

So yes its all bullshit and many fools and money are involved. Any chip that is binned for a specific product is binned for that product, its the most basic thing since forever. I don't need to click links for that. Any time Nvidia thinks a specific die may go into another product, they actually tell you. We've had our share of GP104 movements for example with Pascal. We had A-dies with Turing. In neither case though did those qualify for a product that performed differently from the similar named one with a different chip inside. Even the so called remarkable Turing FE's needed a higher power target to 'be better'. Nuff said.
Oh riiiiiight, I can agree on the article, I mis understood you sorry..

Yay @londiste I was talking bins for pro /enterprise not over clocking , it depends on the cooling too but none of that matters to enterprise, just performance ,TCO.. .
Posted on Reply
#63
Berfs1
R-T-BI just can't see the watts on Samsung 7nm being worse than Turing, which is about what it'd need to matter I think.
Performance per watt is what matters. Not just watts.
AssimilatorWhat exactly does Quadro have to do with any of this?

And of course there's binning. It's just that it won't matter to most consumers. Unless they want to pay more for a better-binned card. Which NVIDIA is perfectly within its rights to charge them extra for.

What will happen is that, since NVIDIA won't be able to sell 10% of its best-binned chips at a higher price (maybe 1%), most of them will end up going into the ordinary consumer pool, and the silicon lottery will work its democratic nature as always. In other words, it'll be just like Turing, and Pascal, and every series before...
Because Quadros are made for reliability (as well as teslas), because reliability is more important in the workstation environment than in consumer cards. Having a graphics card that lasts 1 year longer is better than having it perform 2% faster for the gaming market. That is why workstation/server parts get the higher binned chips, because not only they are the more expensive market, they will care if the server lasts another entire year rather than just 1% more performance.
Vayra86Nobody is denying binning... I am denying the need for this whole article to exist. What's being reported here? What's the point? This is like saying 'Cars run on petrol... but some run on electricity or gas'. Well, thanks for the update :D
Well, TechPowerUp has articles for NASes, so it is clear that TPU is NOT just a news publisher just for regular gamers, but for various different segments in the computer industry as well, overclockers included.
Posted on Reply
#64
EarthDog
AssimilatorBuy a better PSU or don't buy an RTX 3000 card.
This. Undervolting is shooting yourself in the foot...
Berfs1Well, TechPowerUp has articles for NASes, so it is clear that TPU is NOT just a news publisher just for regular gamers, but for various different segments in the computer industry as well, overclockers included.
well, that went over your head.
Posted on Reply
#65
R-T-B
Berfs1Performance per watt is what matters. Not just watts.
That's what I mean.
Posted on Reply
#66
BoboOOZ
Soooo, if yields are so damned good, where the hell are those graphic cards? Only 50 global retailers got them?
Posted on Reply
Add your own comment
Dec 23rd, 2024 01:14 EST change timezone

New Forum Posts

Popular Reviews

Controversial News Posts