He's measuring the temps with a compunurse, but he doesn't specify how. I'd assume an under-the-cpu probe was setup, and so the core temps are probably 5 deg C higher than he indicated.
Also, he only ran the load program for 12 minutes, and IMHO, that's just not enough time to properly measure a temp. You can see it in the graph, all he shows is the temp rise, not where the temp stabilizes. A stable temp for a one minute time period just doesn't cut it for me. Bad review.
Don't beat yourself up, I'm sue Joe would pound on him for such a lousy review. Your temps are achieved, where this rig probably does at least 5 degrees worse than what is stated.
|