Hacker News Reader

Comment by pc86 | Hacker News Reader

I have to wonder why this is done. I know it must make sense or it wouldn't be done, I just don't understand it.

If you're intentionally disabling functionality in order to sell at a lower cost, you're not actually saving any money because you still have to manufacture the thing. It also (I assume) opens up a risk to someone finding out how to "jailbreak" the extra cores and now you can buy an i7 for the price of an i3. Is the cost of having three different manufacturing processes so large that it's not worth switching? Is the extra revenue from having three different versions of the same physical chip enough to justify the jailbreak risk?

View on HN

This is done with cores/memory banks that didn’t pass QC. For example, a 6 core CPU and an 8 core CPU might have the same die as a 12 core CPU, but 6/4 cores, respectively, were defective, so they get disabled. I don’t think they’re crippling fully functional hardware.

See here: http://isca09.cs.columbia.edu/pres/09.pdf

Also here: https://www.anandtech.com/show/2721

“When AMD produces a Phenom II die if part of the L3 is bad, it gets disabled and is sold as an 800 series chip. If one of the cores is bad, it gets disabled and is sold as a 700 series chip. If everything is in working order, then we've got a 900.”

Specifically regarding Phenom II, I have a 550 Black Edition still plugging away, serving different roles over the years, and I was able to successfully unlock and run the two locked-out cores (via a BIOS option). It's never skipped a beat at stock clock. It could be that there was an oversupply of quad-cores, or perhaps (since it was a Black Edition part marketed to overclockers) the extra cores failed when overclocked. I know I wasn't able to have both overclock and four cores, but I considered the extra cores more important, since it was already a reasonably fast chip for its day.

It's likely the latter (that it couldn't work when overclocked with all cores). The market for those is to allow overclocking, so if it can't do any overclocking with all cores AMD likely wouldn't want to sell it as a 4-core Black Edition, since it'd probably just get returned.

> ...but 6/4 cores, respectively, were defective, so they get disabled. I don’t think they’re crippling fully functional hardware.

Hmm, but what if 3 cores are defective? If that can happen(?), then it seems one extra functional core is disabled to get to an even core number.

Apple's M1 GPUs are the first where I've seen the choice between 7 and 8 cores (as opposed to 6/8 or 4/8).

I imagine there is some trade off to be made between increasingly surgical disabling of components and avoiding a menagerie of franken-SKUs. Presumably the fault rate is low enough that tolerating a single GPU core drop takes care of enough imperfect parts.

Perhaps there is fault tolerance hidden elsewhere, e.g. the neural engine might have 17 physical cores and one is always disabled. Although this seems unlikely as it would probably waste more silicon than it would save.

> Hmm, but what if 3 cores are defective?

It gets sold as a coaster

More likely, a keychain:

https://www.amazon.com/Keychain-Ryzen-Threadripper-Computer-...

https://www.amazon.com/Keychain-Intel-Core-Computer-Chain/dp...

Kind of funny that Amazon recommends it be bought with thermal paste

A coaster that allows you to play Doom.

This is done because your production yield is not 100%. So instead of throwing away every produced component which doesn't achieve the target of your 100%-product, you "soft-lock" the components with 80~99% performance into a 80%-product category, and the ones with 60~80% into a 60%-product. This way you increase the total yield-rate, and produce less waste. The counter-intuitive waste happens when the demand for the 60%-product is exceeding your "natural" supply of 60%-output, so you have to start to "soft-lock" some of your "80%-product" production to the 60%-grade to fulfill demand...

Are you guys sure? I think manufacturing has nothing to do with it.

The real reason IMHO, is to have a larger range of product prices so you can cater for specific audiences.

It seems people are confusing cost with price. Those two things are orthogonal.

This tends to be the case later on in a product's production run, as the manufacturer has fine tuned the process and worked out most of the kinks, the pass-rate of finished items increases.

At this point, yes they may lock down perfectly good high end CPUs to a midrange model spec to meet a production quota.

But do these production defects really meet the demand of the lower tiers? Also how is it possible to predict the number of defects in advance so that they can make useful promises to distributors?

Well eventually as yields improve you start handicapping perfectly valid chips to main market segmentation.

I cannot say this for certain in CPUs but I know in other electronics with PCBs that this is how it is done. Sometimes lower-end SKUs are made by opening a higher-end one and cutting a wire or a trace.

I'm curious about this as well. It seems inevitable that some batches will be "too good" to satisfy demand of low end chips.

Either they just accept the fluctuations in order to maximize output of high end chips, or they would have to cripple fully functional ones to maintain a predictable supply. Interesting business.

It's not primarily about using defective chips (but that's a nice side effect). As a process becomes mature, yield rates become very high and there wouldn't be enough defective chips to meet demand for a lower tier, so good chips are binned into those tiers anyway.

The primary purpose is market segmentation: extracting value from customers who would pay more while not giving up sales to more price sensitive clients who nevertheless pay more than the marginal cost of production.

That makes sense, thanks. I wonder if it would be possible to de-bin one of the lower end ones, assuming it is a binned version of a fully functional higher tier chip. Or perhaps they completely destroy the offlined cores/features.

The term for this is "binning", and the explanation is wholly innocent. Manufacturing silicon chips is not an exact process, and there will be some rate of defects.

After manufacture, they test the individual components of their chips. These chips are designed in such a way that once they identify parts of a chip that are defective, they can disconnect that part of the chip and the others still work. (I believe they physically cut stuff with lasers but my knowledge is out of date). This process can also includes "burning in" information on the chip itself, like setting bits in on-die ROMs, so that if your OS asks your CPU for its model number it can respond appropriately.

Interesting side note: The same thing happens when manufacturing even basic electronic components like resistors. All the resistors that are within 1% of the target resistance get sold as "±1%" resistors, which means it's pretty likely that if you buy the cheaper "±5%" resistors and test them, you'll find two clusters around -5% and +5% and very few at the target value.

> The same thing happens when manufacturing even basic electronic components like resistors.

EEVblog did some tests[1][2] some time ago on ±1% resistors, and found that while his samples were fairly Gaussian and within the spec, the ones from his second batch were consistently low. That is, none were above the rated value.

So yeah, don't assume a perfect Gaussian distribution when using resistors.

[1]: https://www.youtube.com/watch?v=1WAhTdWErrU

[2]: https://www.youtube.com/watch?v=kSmiDzbVt_U

That's because there is a huge variance in the quality of chips produced because the process isn't 100% precise.

So the best chips which have the least errors in the manufacturing process are sold as top tier. The ones which have more mistakes in them get their defective parts disabled and then get sold as lower tier ones

It's called 'market segmentation'. It's why there are different brands of soap powder from the same manufacturer even though they are all essentially the same.

Yep. I don't think it has anything to do with manufacture issues.

> I don't think...

Instead of assuming, it's easy enough to confirm that CPU binning is real.

I think of you Google price discrimination or similar economic terms you get some explanations for this.

If you just have one price, you cut out people who can't afford it and people who can afford to pay more get away with more of the surplus.

If you have several prices and create just enough difference in the product that it doesn't change the expense much, you can suck dry every class of user.

Bit of an MBA trick.

"suck dry" is excessive editorializing for a practice that make sit possible for a market to exist without, well, "sucking the manufacturer dry".

Exactly, it is not an economics trick for if companies could only supply a certain level of the market it might not even be profitable to produce. It is just economics...

We are talking about chips here so not the same bottle of soap with different labels, I'd call that a trick.

Reply on news.ycombinator.com