Hardware

Gigabyte Ends the Year with a Bang: Kills AMD FX-8350 CPU

UPDATE January 2nd, 2013 – We’ve been in contact with GigaByte’s Taiwan HQ and its North American subsidiary to work on the issue of the failed component. The motherboard in question served as a review platform for several AMD processors and we’re sending it back to Taiwan in order to see what components failed. According to GigaByte, the company updates its components on the motherboard to ensure reliability. Our version was 2.0, and it featured new electronic components.

Original Article
When people talk about ending the year with a bang, they usually talk about festivities involving beautiful fireworks. In the days after Christmas we experienced our own episode of ending the year with a bang in quite the unexpected way.

If you cannot procure good high quality fireworks, Gigabyte GA-990FXA-UD3 might do the trick
Gigabyte GA-990FXA-UD3 has a great color coding for a Black/Red or Black/Green PC, but unfortunately did not serv us well during the FX-8350 test.

What we are talking about is an unfortunate incident that happened during one of our benchmark sessions. We were running benchmarks of AMD FX-8350 "Vishera" processor on a Gigabyte GA-990FXA-UD3 which ended abruptly when a MOSFET on the board decided to fail. We had a bad feeling about it when we noticed the smell of burnt plastic and removing the VRM heatsink confirmed our worst assumptions as our pictures show.

The blown MOSFET.

The blown MOSFET.

Thermal pad on the VRM heatsink clearly shows the aftermath of the blown MOSFET.

Thermal pad on the VRM heatsink clearly shows the aftermath of the blown MOSFET.

We’d like to note that the incident happened when running benchmarks and stress tests at stock voltage and clock speed. During our testing, we found that the VRM heatsink of the board got very hot, so hot that you can’t rest your fingers on it for more than a moment. We don’t know whether the part was getting too hot or it was merely a failed part but we would conclude that the VRM heatsink doesn’t look very efficient from a cooling standpoint. It should be improved to ensure cooler and safer operating temperatures.

Before running our benchmark suite at stock settings, we did some overclocking tests, which might have contributed to the MOSFET failing. We followed AMD’s guidelines from the Reviewer’s Guide for safe operation under overclocking (5GHz with some +0.175V on the chip), where the chip passed all the stability tests with flying colors.

Our power consumption measurements showed that for overclocking Vishera, you better get top notch cooling. Power consumption went from some 260-280W at stock under full load to slightly over 400W in an overclocked configuration (power consumption figures are measured at the wall outlet). We continued with our stock testing when the unfortunate incident happened. Further investigation of the incident showed that the CPU went the way of dodo birds as well. Our secondary AM3+ board, which happily boots our FX-8150 "Zambezi" processor, the FX-8350 "Vishera" wouldn’t POST anymore. Other components such as RAM, graphics card and PSU were not harmed in the process.

This unfortunate event only goes to show that no hardware is bullet proof. Unfortunately, in our dealings with Gigabyte EMEA staff, we were unable to secure a new part "as they do not ship to Vienna, Austria." We hope to be able to bring you the in-depth result of our month-long review soon, depending on procuring a new CPU and the motherboard.