Intel Atom chips have been dying for at least 18 months

>The flaw in Intel's Atom C2000 family of chips has been vexing Intel's hardware customers for at least a year and a half, according to a source at one affected supplier, but it wasn't immediately obvious that Intel's silicon was to blame.

>The well-placed insider, who spoke to The Register on condition of anonymity, said the problem – which results in bricked systems – became apparent to engineers at product makers when the return rate on gear spiked about 18 months ago.

>It took additional time to figure out that Intel's hardware was responsible. The vendor in question worked with Intel engineers during this period to replace affected units under manufacturer warranty. Then about two months ago, Intel acknowledged the issue was with its chip blueprints, the source said.

theregister.co.uk/2017/02/07/intel_atom_failures_go_back_18_months/

Other urls found in this thread:

ark.intel.com/products/codename/63508/Denverton
en.wikipedia.org/wiki/Goldmont
en.wikipedia.org/wiki/List_of_Intel_Celeron_microprocessors
ark.intel.com/products/codename/54859/Avoton
en.wikipedia.org/wiki/Silvermont
computerworld.com/article/2497201/windows-pcs/intel-shifting-celeron--pentium-chips-to-atom-architecture.html
twitter.com/NSFWRedditVideo

>Networking gear

Doesn't effect my meme chink consumerist tablets which use Z series so I don't care :^)

The Apple iPhone 7 Plus doesn't have this problem.

A lot of these went into embedded systems, and they cannot be replaced. Expect intel to get sued big over this.

Considered getting the C2750D4I for 4 slots on ITX, glad I didn't in the end

INTEL JEW SHILLS ON A SUICIDE WATCH!

I almost went for a C2750 too, good thing I decided to wait for Zen

This would explain why my asrock c2550d4i motherboard keeps fucking dying.

The iPhone is not in competition with hardware that uses the Atom C2000 chips. It was designed for embedded platforms, not phones.

Expect a BIOS update probably.
Unless the article says what it actually is.

>problem with the electrical circuits
>expect a bios update
how retarded are you?

Fuck me I just RMA'd my C2550D4I. I thought it was just typical AsRock shit.

ark.intel.com/products/codename/63508/Denverton

Intel is replacing C2000 with C3000 soon, it's not an issue

>x is flawed.
>just buy the next one!
Apple logic.

Yes let me quick recall my satellite from orbit for a quickie CPU upgrade and we'll be off and away.

Oh that's right, people don't typically put Intel CPUs into satellites in orbit. Because that'd be fucking retarded because Intel can't even nail down and produce even just one long-term, stable X86 / AMD64 CPU.

INTEL JEWS BTFO AGAIN

>only server cpus
ok

How much you anons want to bet it's something to do with their fucking built-in fuses.

Those fuckers could have been cooked up with the wrong memes and are popping now...

Does Intel even have any radiation hardened CPUs?

All of them.

Well this looks like the self-destruct function that intel was working on before Zen made them actually give a shit

Intel has their own fucking particle accelerator- they test ALL their semiconductor products in particle accelerators.

It's not average day radiation that's the problem, it's cosmic rays from space that come down with extremely high energys and fuck up processors.

my ass
bla bla marketing

They actually do, dipshit.
So does AMD.
So does ARM.
Everyone particle accelerate tests their processors.
you'd be a moron not to.

So you're telling me that every day Intel CPUs have the same radiation hardening that makes a difference between a $100 processor and a $200000 processor? Bullshit.

sure user

DELET

?super low binned failed celerons aren't selling
Oh no, how will intel survive

>aren't selling
more like, failing in the field

>implying 8 core Celerons even exist

>failing in the field
>Super low binned failed celerons are failing
Oh wow, nobody saw that coming

No.

Doesn't mean that they don't have radiation hardening.
Check your reading comprehension nigger.

I'm not doubting that they have some form of radiation hardening but I was clearly talking about radiation hardening for use in space. Obviously you're the one that needs to work on his reading comprehension.

So Intel's just fine with selling defective products, good to know.

Space-hard CPUs are few and far between and typically use expensive silicon on sapphire technology. I'm not aware of any space-hard Intel CPUs.

>fine with selling defective products
All binned products are defective products.

Still waiting for you to show me that 8 core Celeron that becomes these Atoms

en.wikipedia.org/wiki/Goldmont

>binned
Was Intel selling these defective CPUs without telling the customer they were likely to fail?

No, of course not.

Anyway I don't think you'll find much support for your claim that all CPUs that ship with cores disabled are 'defective' inside Intel. It's a yield-increasing measure.

>buy cheap as shit low power CPUs
>surprised they don't last as long as more expensive parts

Oh wow $20 parts die when they're out of warranty

There isn't a single 8 core processor listed. Here's a list of Celerons, notice that there isn't a single 8 core processor on that page either.
en.wikipedia.org/wiki/List_of_Intel_Celeron_microprocessors

>8 core Atoms
>Literally 4+ year old parts long out of warranty failing

Oh wow, how surprising.

ark.intel.com/products/codename/54859/Avoton

All these server Atom CPUs are only sold as Atom, not Celeron

Denverton uses Goldmont cores, up to 16 cores while Avoton maxes out at 8

>claims the processors are low binned Celerons
>there isn't even a Celeron in existence that could become one of these
Just admit you were wrong already.

see

What does that have to do with the fact that there are no 8 core Celerons?

en.wikipedia.org/wiki/Silvermont

They use the same silicon and process

Irrelevant. You can't take a low end quad core CPU and magically double the core count.

Maybe it's the other way round, but they're all budget tier parts is the point.

And my point is that it's impossible for an 8 core Atom to be a low binned Celeron. Sure, the Atoms may be low end parts but they were never Celerons.

Celerons are failed Pentiums, which are failed Atom server processors.

No such thing as an 8 core Pentium either. These processors were destined to be Atoms from the start.

Just look at this shit and figure out the binning.....
8/4/2 core silvermonts that make the grade become Atoms, Pentiums or Celerons depending on how defective they are....

>the two core Atoms are literally failed celerons
>The autstic state of this retard who doesn't understand product binning

>C2350 aren't FAILED CELERONS REEEE
>literally downclocked celeron that sells for $2 less

So where are these 8 core Pentium and Celeron processors that become Atoms when binned low enough? Oh that's right, they don't fucking exist. Also take those ellipses and fuck off, you retarded twat.

Remember goys, mobile processors dont have sockets. They are soldered directly to you (unique) motherboard.

tldr: GET WREKED GOYIM, NOW BUY NEW STUFF

>still focusing on muh 8 core
Hey retard, some atoms are failed celerons. Nobody said that all of them are.

Hey dumbass, I never said the lower core count ones aren't. I'm clearly talking specifically about the 8 core models.

>I'm talking specifically about the 8 core models
And nobody said that 8 core models are celerons.

Sure thing bud

see

I was obviously talking about 8 core processors.

The ones that are failing are literally C2338s you stupid fuck. They're literally failed celerons

the top of the line Atoms aren't the ones failing you dumbshit.

Yea my Baytrail tablets that were like 80 bucks or less are still going strong. They run Crysis have a full size USB 3 Type A Micro HDMI out Micro SD and sufficient eMMc 32GB. They run Win8 or Win10 which I eventually tossed Android x86 on with grub. I also have a Asus ZE551ML with a similar chipset and it's comfy as hell. Great with emulators too.

The 8 core ones are okay?

>different TDPs
>celerons dont suppot ECC
>celerons have on board graphics
>atoms have higher turbo modes
>same chips but atoms are lower binned
i'm not that user but how retarded are you

Fucking Intel we have bought Synology NAS with C2538 processor not so cheap 2month ago, and this shit intel processor with long term support for 7Years dying in 18months, Great. Never ever. In future AMD or ARM.

>Fucking Intel we have bought Synology NAS with C2538 processor
You paid $2200 for a RS2416RP+ when you could have just bought some supermicro parts and installed windows or linux for a fraction of the price. How retarded are you?

Yeah with no time to build things from parts and tunning SW on this we just bought complete solution. Arrrgh

>no time
>to spend an hour assembling parts
>and installing a couple windows features
>instead I have time to migrate ~40TB of data off it in a year and a half over 1GbE
You could have had 10GbE too if you built it yourself

Is their a front end software as simple as synology?

>different TDPs
nigger you blind, they're literally 2 watts apart and downclocked

>TDP is lower because clockspeed is lower
>I have never heard of underclocking
This is how retarded you are.

>C2338
>7w TDP
>1.7ghz
>2 cores
>1mb level L2 cache $60

vs
>10w TDP
>2.41 Ghz
>2 cores
>1mb L2 Cache $72

Totally not an underclocked binned celeron m8

Seems so, the article only mentions the literally piss poor parts specifically the C2338
>ASROCK board dies
>It's intel's fault

>Is their a front end software as simple as synology?
Yes. If you run ESXi on it and setup a failover cluster you wont incur downtime when you install windows updates either.

see you retards. the atom is clocked higher yet you keep on claiming it is a lower binned celeron

>atom is clocked higher
Only the top of the line ones you stupid fuck
The 'atoms' you mentioned are communications boards not sold to the public.

>clocked higher
>The TDP is literally double because it's two pentiums sewn together
You really are a moron you know.

Hey dipshit 4 core pentiums are literally Top of the line Atoms with defective cores being turned into IGPUs.]

>the top of the line ones
>are lower binned celerons which dont have specs nearly as nice

>its a pentium now not a celeron
>because pentiums are dual socket
>i dont understand what cache coherency is

>The vendor in question
aka cisco

>top of the line ones are lower binned celerons
Who said this?
People are saying that the defective ones are literally low binned Celerons.

See
>c2350
>Literally an underclocked celeron j1750
>and costs $2 less

>It's a pentium now not a celeron
THEY'RE ALL THE SAME FUCKING MICROARCHITECURE AND PROCESS YOU STUPID PRICK. THEY'RE ALL BINNED VERSIONS OF THE SAME FUCKING THING JESUS FUCK YOU ARE RETARDED.

>ZE551ML
MAH.
FUCKING.
NUIGGGGGAAAA

>8 core atom
>Hey guys some of the cores are defective.
>It's okay we'll just cut the silicone and tag an IGPU on it and sell it as a pentium

Welcome to the world of binning.

>Hey guys even more of the cores are defective
>It's okay we'll just disable the 2 cores and sell it as a Celeron

>Hey guys this one can't reach the required clockspeed.
>It's okay we'll just underclock and disable the turbo and sell it to retarded cheap server machines

Yeah the fucked atoms are literally parts discarded 2/3 times

computerworld.com/article/2497201/windows-pcs/intel-shifting-celeron--pentium-chips-to-atom-architecture.html

After Ivy Bridge, all Pentiums and Celerons are basically different variations of Atoms and vice versa. They're literally the same product that's binned for different markets.

Fibre optic cables are too easy to break though

like pottery

This isn't uncommon. Intel has been running scared because of ARMs domination.

They've fired a lot of their quality control people in hopes of accelerating delivery to market

Basically, be careful when buying garbage from Intel and check the erratas for any stepping you consider purchasing.

...

...

You have no idea what you're talking about, manlet.
Shut your mouth.

ARM domination is hilarious. And skylake had half the amount of bugs haswell did

this is a big shitup, especially with AMD coming back
Shittel must be in panic mode about now

One of the fix purported by one of the big networking vendors is a reduce clock rate, which could be possible via BIOS updates, but god knows what problems that could cause with an SoC.

>The ones that are failing are literally C2338s you stupid fuck
And the C2550/2558
And the C2750/2758
Otherwise known as ALL OF THE FUCKING AVOTON/RANGELEYS

Also, the C2750, C2550, and C2358 use a different die than the consumer Pentium/Celeron SoCs. The dies are significantly larger and has more embedded controllers in the die. A C2358 has up to 8 lanes of PCIe Gen 2.0 along with an embedded NIC with support for 4 1GbE ports.
Not to mention that the Avoton/Rangeley uses a different FCBGA (FCBGA1283) than the Pentium J2900/CeleronJ1900/CeleronJ1800 (FCBGA1170).

Now hang yourself.