this is also why servers and chips that run for a long long time (think servers that aren’t decommissioned for years) don’t run such high voltages and clock speeds.

What I would want to know is what happens to these good binned chips when you lower their voltages and frequencies and run them at a more reasonable 3.5GHz or something for a long time. Is the price / power / performance ratio better than a server chip or is it worse ? Would love some data on that. But no one is buying these for it and server chips have different things going on for them like being able to replace the chip for years etc. still would be interesting to find out if these the best ‘quality’ chips out there

> What I would want to know is what happens to these good binned chips when you lower their voltages and frequencies and run them at a more reasonable 3.5GHz or something for a long time

I'm not sure about intel chips, but amd's last couple generations of chips have an "eco mode" which underclocks the chip for you. You get about 85-100% of the performance while consuming 60% as much power. The chips probably last way longer like that too. NVidia's GPUs have similar options. 3090/4090 cards can have a totally reasonable power budget if you're happy to lose ~10% of your framerate.

Arstechnica added eco mode to their benchmarks:

https://arstechnica.com/gadgets/2023/03/ryzen-7950x3d-review...

Thanks. The 7950X doesnt have a non X variant. So it seems the non X variant chips are basically doing the same - giving you lower clocks for lower wattage with an option to tinker at your own risk for pushing the chip as those are not as good as the X variants.

I run 50mv undervolted 12500 which gives me about 20% lower consumption.

Where are you getting that ECO mode cause a 15% performance drop?

In my pretty unscientific testing, I saw ECO mode have zero impact on single core workloads and ~4% on multicore with a big drop in cpu temperature.

If I am reading the “Gaming CPU” chart correctly from your link, the AMD 170w performance was 114.8 FPS, but the 105w ECO performance was 111.4 FPS. Effectively nothing for a huge drop in power.

Was this on a laptop or desktop? If a huge drop in power results in no meaningful drop in performance I would say that is a sign that previously you were hitting a point where heat was more of a problem than power.

My testing was on a desktop with a way over specced cooler. Regardless, the above link shows that they were getting huge power drops for basically no loss in performance.

With makes sense to me. We have seemingly run out of cheap architectural wins. CPUs and GPUs keep cheating by amping up the power draw to improve the performance numbers for vanishingly small returns. To get an extra 5% performance on these chips can require 10s of watts.

> Where are you getting that ECO mode cause a 15% performance drop?

It depends on the test. In that link I posted, 3dmark’s cpu test shows eco mode drop the score from 19000 to 14000. Which I guess is more like a 25% perf drop. In other tests, performance increased in eco mode.

It seems like it depends a lot on the workload.

I design chips in similar technologies. What these binned parts are fast-fast parts, meaning both PMOS and NMOS transistors have lower threshold voltage than typical due to stochastic process od production or intentional skewing of the process during fabrication. You can run them at faster clock speeds. If you stack higher voltage on top you can run quite a bit faster. For a while.. The transistors age. Aging is an exponential function of voltage level and temperature, and lunear function of clock frequency. With aging the device threshold voltages increase rapidly, so the devices get slower. On top of this there is electromigration (EM). With EM the interconnect resistance increases. These effects combined, you have a horrible product lifetime at those cobditions. After failung to work at 5-6 GHz at 1.4V, it would most likely still work at a lower clock frequecy, because in essence it becomes a slower chip.

Now the answer your question: A fast binned part operating at nominal conditions will perform exactly the same as the same sdries typical chip. The power consumption would be like 10% or so higher due to higher dynamic power and leakage. The lifetime would be muuuch more longer than a typical device though. So, in short, wouldn't be outperforming anything.

I would suggest not spreading FUD unless you can cite hard evidence.

Intel warrants their processors for 3 years if they are operated according to spec, which includes supplying 1.3~1.4V as far as I'm aware.

From what I can tell, the linked thread and article concern CPUs that were either defective out of the box (and thus under warranty, regardless usage) or were driven out of spec by mobo overclocking by default (which would void warranty).

It's not FUD to point to real problems that are occurring in the field. A processor that's unstable is still unstable even if it's covered by warranty. And to the extent that these real issues are being caused by aggressive motherboard defaults, Intel is not absolved: they have huge leverage over their motherboard partners, and are seemingly failing to use it to ensure users get a good, stable experience out of the box (likely because Intel cares more about high scores on popular benchmarks than about subtle stability issues in this market segment).

Yes, real problems stemming from factory-defective products or products that were driven out of spec. The article linked even admits they (author or publisher) don't have enough evidence to pinpoint what the problem is; the best they could do was the aforementioned. Nowhere do they say "Driving 1.4V is killing CPUs." or something similar; just potential workarounds like reducing clock multipliers below spec and configuring mobos to enforce Intel's power limits.

Drive known-good products according to published specifications at load for statistically significant durations. If the results are that the majority of products fail to perform as warranted, then we can talk about how Intel (and I guess AMD) are driving their products to the point of failure.

Otherwise in the absence of such data, I'm going to look at the silent majority satisfied with their purchases and infer that the products concerned are working fine.

> If the results are that the majority of products fail to perform as warranted, then we can talk about how Intel (and I guess AMD) are driving their products to the point of failure.

That's a stupidly high bar. Recalls and class-action lawsuits don't need to be justified by failure rates as high as 50%, and I'm merely discussing that there are signs of trouble, not demanding a recall or other serious action from Intel. Intel's recent top of the line desktop chips are misbehaving in a way that is genuinely noteworthy, even if we don't have the impact solidly quantified and don't have a smoking gun. It's worth discussing, and worth keeping an eye out for similar issues from other chips that are being pushed to similar extremes.

>Intel's recent top of the line desktop chips are misbehaving in a way that is genuinely noteworthy, even if we don't have the impact solidly quantified and don't have a smoking gun.

And all I am asking is for you to cite proper evidence for your claim. The article you linked does not say driving 1.4V is damaging the CPUs, it's actually explicit that the cause is unknown. Speaking more broadly, most people who have bought the CPUs concerned have had no problems (or at least do not voice such concerns).

To reiterate, I am asking you to cite evidence for your claim that "Intel and AMD are pushing 1.4V into chips built on processes that work best at and below 1V. And the result isn't a stable, long-lasting chip." If you can't or won't, this is just FUD.

If the instability is something that develops over time as a genuine change in behavior of the chip, and not merely an artifact due to the evidence of instability taking time to pile up, then the extreme voltages are by far the most plausible culprit. And if on the other hand these chips are slightly unstable out of the box, despite the high voltages required to hit these peak frequencies and record-setting benchmark scores, it suggests that the clock speeds are being pushed too far.

Either way, the high operating voltages compared to what we see in laptop and server CPUs (and GPUs for any market segment) is worth raising an eyebrow. At a minimum, it's a symptom of the desperation Intel and AMD have for perennially leapfrogging each other in ways that are increasingly irrelevant to the average customer and the rest of their product stack.

Would be nice to refer to something that does not say the reason is unknown anyway.

Some of this is modeled over expected life-time / usage. In general, things such as electro migration, self-heating, bit-cell degradation, etc. are modeled either for 3y, 5y or 10y, depending on CPU, skew, and target market. Now, whether the process corner(s), voltage, frequency that was picked to perform this analysis is a good reflection of how the CPU is being actually used/pushed, is a different matter.

This explains a lot about my computer experiences over the last few years. I've been buying absolute top end hardware, and been consistently running into a plethora of absolute weird technical issues unlike any of my previous experiences. I even had to rma a 12900ks directly with Intel after it couldn't run at stock settings bug free.

>If the instability is something that develops over time as a genuine change in behavior of the chip, and not merely an artifact due to the evidence of instability taking time to pile up, then the extreme voltages are by far the most plausible culprit.

That's all fine and dandy, but can you please cite some evidence to support those claims?

This should not be such a farfetched request.

Exactly, a warranty is fine in theory, but I care about machine uptime and where my own time is going.

Often the time to deal with warranty exceeds the cost of a new component, in which case I don’t bother and just try a different brand.

Case in point, had some trouble with WD NVME drives - just tossed them, bought Samsung, and moved on.

Wasn't there an article here on HN more recently? Anyway, check out https://semiengineering.com/the-rising-price-of-power-in-chi... which also describes a few issues with actual chip aging. That may explain why Intel for example gives only 3 years of warranty, despite the fact that people still believe in "that is solid state electronics, it will last a lifetime!"

Also, the overall cost of design and production is going up for years (>10 years already or so), which is why only these few fabs are left that run these advanced processes and they need massive scale to stay profitable. This "going up" is btw. non-linear, more like exponential and that also applies to power consumption of the resulting chips if you want to eek out "a bit more".

I thought the 3 year thing was standard for many years, pretty much as long as there have been retail-boxed CPUs.

I think the logistics of a long term CPU warranty might be difficult too; aside from a few perennial, typically embedded market, products, manufacturers want to close out product lines and keeping warranty spares for decades becomes a burden.

Imagine the fracas at the RMA office: "He says he doesn't WANT a new 14900KS, he wants a 50MHz 486SX2 to replace the one that blew out, and he has the state attorney general on line 2."

Most PCs are operated for way longer than 3 years, but still are there any reports of dead CPUs that were operated in spec? With millions of devices out there, you'd expect some to pop up, but I only ever hear of dead HDDs or defective RAM.

Had a 6900k bail out on my after 4-ish years. Got a new one from extended warranty.

I recently had mother board die on me, I think at age of 4 years and bit... Everything else is still working fine. So that might be in the pile of things that could act up in some cases.

There was no FUD there at all. Grow up.

It is FUD to state that 1.4V damages CPUs without citing evidence to support that statement.

Personally, I have a 14700K in my desktop and my laptop has a 12700H. Both routinely push 1.3~1.4V under load operating according to spec. If that is causing damage, I certainly would like to know and I asked for citation. I see none after prodding, so as far as I'm concerned it's FUD.