There a serious problem with the Atom C2000 series processors from Intel. After certain amount of time the internal clock generator of the CPU will degrade substantially causing defects in the produc... Intel Atom C2000 chips are bricking products
Man I looked at a small c2000 integrated system with 8 cores and now I'm glad I went against it. Those suckers are not cheap.
in article it said the chip released on 2013 and it failing after 18months (so its 1.5years) its said the chip mostly used in NAS / network equipment which should be online most of time if not 24/7 why it detected/reported 3years later? such this should detected sooner because most device is start to failing after 1.5years of usage
If you want many low-power cores, ARM is definitely the way to go, assuming your software and other hardware is compatible. If size isn't an issue, there are some relatively cheap 6c/12t Xeons that are low-power.
Most (if not all) Atom CPUs are soldered to their motherboards. Usually when this is the case, manufacturers don't tend to use the latest and greatest. So even though this is a 2013 CPU, there are likely products released in 2015-2016 that are using these. Generally speaking, industrial class hardware is a bit outdated. Ironically, because the outdated stuff is supposed to be considered proven stable.
It probably also takes time to pinpoint the exact reason the hardware failed, and what this article is telling about is extremely exact. Intel's own investigation must have taken months as well (and they are still refusing to comment directly, by the looks of it).
That doesn't change the fact there was an explicit hardware defect. When something threatens the integrity of a company's functionality, getting yourself up and running (in other words, replacing the defective part) as quickly as possible is more important than figuring out what's wrong. So, it would make sense to say "there is a problem with these CPUs where they will fail after a certain amount of time" and leave it at that. These CPUs are BGAs, so there's nothing the customer can really do about the problem anyway and therefore the culprit of the problem is irrelevant.
Is anything from Intel cheap for what it is? How the hell did Intel manage to screw this up... the clock generator degrading? Seriously?
I don't think the specific functional block matters. It could have been any particular segment. Failures happen. Given the design cycle time to RTM and that it took well over a year to exhibit failures. I don't see this as particularly unusual except in scope.
If anyone 'member the original P67 chipset release, there was a similar problem where the SATA ports would degrade.
There's lots of examples...sometimes they just fade away as 'errata'. Folks should look that term up in regards to processors.
With the SATA motherboard issue they offered a free replacement B3 Revision if memory serves correct. At least Asus did anyway. B3 issue was discovered alot earlier on however so will be interesting to see how this is handled.
They did a recall. It ended up costing them close to a billion. Like the other issue, it isn't a functional issue, it's a statistical issue. A certain amount will fail over time.