Only 2% of the same size was overclocked with the caveat that only 477,464 machines within the sample could have their proper clock speed identified.
Overclocked is defined by being 5% outside of rated speeds.
We have divided the analysis between two CPU vendors, labeled “Vendor A” and “Vendor B.” The table shows that CPUs from Vendor A are nearly 20x as likely to crash a machine during the 8 month observation period when they are overclocked, and CPUs from Vendor B are over 4x as likely. After a failure occurs, all machines, irrespective of CPU vendor or overclocking, are significantly more likely to crash from additional machine check exceptions.
The data implies that AMD and Intel also have a substantial difference in the manufacturing quality of their chips. Who is who in this study is an interesting guess.
It also implies that overclocking will sooner rather than later bite you in the ass.
As for OEM vs White Box.
We identify a machine as brand name if it comes from one of the top 20 OEM computer manufacturers as measured by worldwide sales volume. To avoid conflation with other factors, we remove overclocked machines and laptops from our analysis.
So overclocking did not taint the result that OEM is more stable than white box. Anything not one of the top 20 OEMs is a white box, so DIY boxes do fall into the white box category.Edit:
As one answer to my own musings. Most OEMs slightly underclock their machines.
Therefore, we further partitioned the non-overclocked machines into underclocked machines, which run below their rated frequency (65% of machines), and rated machines, which run at or no more than 0.5% above their rated frequency (32% of machines). As shown in Figure 5, underclocked machines are between 39% and 80% less likely to crash during the 8 month observation period than machines with CPUs running at their rated frequency.
A small change can have a rather large payback in stability.
"Welcome back my friends to the show that never ends. We're so glad you could attend. Come inside! Come inside!"