Intel's Atom C2xxx processors may just make like a banana and split

Subject: General Tech | February 7, 2017 - 01:31 PM |
Tagged: Intel, c2000, Avoton

"System May Experience Inability to Boot or May Cease Operation" is not the errata note you want to read, but for those running devices powered by an Intel Avoton C2xxx family Atom processor it is something to pay attention to.  The Low Pin Count bus clock may stop functioning permanently after the chip has been in service for a time, rendering the device non-functional.  Intel had little to say about the issue when approached by The Register but did state that there is a board level workaround available to resolve the issue.

The Avoton famliy of chips were released in 2013 and were designed to compete against ARM's new low powered server chips.  The flaw is likely responsible for the issues with Cisco routers that have been reported on recently; the chip can also be found in the Synology DS1815+ and some Dell server products.  It will be interesting to see how Intel responds to this issue, they have a history of reluctance discussing flaws in their product's architecture.

View Full Size

"Intel's Atom C2000 processor family has a fault that effectively bricks devices, costing the company a significant amount of money to correct. But the semiconductor giant won't disclosed precisely how many chips are affected nor which products are at risk."

Here is some more Tech News from around the web:

Tech Talk

Source: The Register

February 7, 2017 | 03:00 PM - Posted by Anonymous (not verified)

That's going to be a very NASty surprise for many for those types of devices based on these Avoton/ATOM SKUs and any other devices based on these C2xxx series models/steppings that have the problem.

It's going to be one big Atom Cash Bomb for Chipzilla! And a somewhat timed released one at that, around 18+ mo before thing can begin to go boing on those clock pins. Better hope that the warrenty is 3+ years on any devices that where purchased with C2xxx inside.

February 7, 2017 | 11:01 PM - Posted by Anonymous (not verified)

I have found the built-in i350 quad GE on some Supermicro boards has a tendency to randomly encounter PCIe bus errors when the IP function of the interface is set to DHCP. I have yet to completely bisect the issue into hardware or Linux kernel, but recovery requires a full system hard boot; power-on reset, not push the button reset.

I have found some models of Intel dual and quad GE boards can cause networking issues in the onboard PCIe slots of some Supermicro boards.

Now I have to worry about this issue on my Supermicro boards with the C2xxx family of processors. Geez. At least Supermicro claims the boards built with the Rangeley family of processors have a 7 year "lifecycle". I still want to check the warranty terms since the CPU is soldered to these boards.

You would think with all of Intel's years of experience in making CPUs and chipsets they would have gotten stuff like this right years ago. What is especially worrisome is that Intel is claimed in "The Register" article to have a "board level" fix for some systems and an update to the CPU design. That sounds to me like "don't count on a software fix" for this problem...EVER!

February 9, 2017 | 12:42 PM - Posted by JohnGR

Nvidia bumpgate?
Nvidia's fix was to increase the fan speed to make sure that the GPU will fail after the expiration of the warranty period, now I am reading about limiting the voltage to the chip or something like that.

February 9, 2017 | 04:57 PM - Posted by Thedarklord

Rutroh raggy!

This will negatively impact several NAS users, myself included. I have a Synology DS 1515+ which runs on one of these...

February 13, 2017 | 10:36 PM - Posted by Thedarklord

Well S***! Not sure if it's related, but my unit just died today... I purchased the Synology 1515+ from Newegg (new) in June 2016 and I got home today to discover it died Feb 13th (hmm) 2017.

I have no idea if it's related to the SOC dieing but it's uncanny that I read this article then a few days later it dies...

Just created a support ticket...

February 13, 2017 | 10:36 PM - Posted by Thedarklord

Well S***! Not sure if it's related, but my unit just died today... I purchased the Synology 1515+ from Newegg (new) in June 2016 and I got home today to discover it died Feb 13th (hmm) 2017.

I have no idea if it's related to the SOC dieing but it's uncanny that I read this article then a few days later it dies...

Just created a support ticket...

February 10, 2017 | 10:30 AM - Posted by Crabbyprime (not verified)

I have a Synology DS1515+, should I replace it? It works fine right now buy having a ticking time bomb in my home is just a smidgen worrying.

February 15, 2017 | 08:13 PM - Posted by jerky33

UGH, my WD My Cloud DL4100 runs on one of these f#@*ers.
Boo Intel!!!
Thinking about replacing it with a diskless PR4100 before the DL4100 dies on me.
Will need to move all my data off of it and back on, don't look forward to that.