Printable View
I have doubts about the details and consequences of mistakes made in the prime test. Some overclocks that I did have errors in the tests of Prime.
But the system works normal, benchmarking tests, get the best results as increase the frequency of the processor, games usually work and other programs.
So the question is, what are the consequences of a mistake in the test? Because I see no apparent problems.
You're doing tests with the Prime Orthos or with the (more common)? But, either with any of the 2 that is, you know how old they are, can make mistakes in Windows Vista, depending on the configuration, which is not, for example, the micro side, never makes an error, and in the micro primary signing with Vcore of 1.31 V gives error after a few minutes! And even if I increase the Vcore to 1.35 V and boost the FSB and PCI Express on 0.1V, the stress test gives error in the same way!
So you know what I did? Vcore go back to 1.31 V and voltages of the FSB and PCI Express in normal, i.e. increase of 0.0 V! and restart the microwave!
Then the Crisys Warhead, played for 2 hours and it worked fine without any crashes! Now, if I activate the Orthos, after 1 minute longer would mistake!
So, who is the error?
That is relative. Even though older, they tend to be efficient. The calculating Prime95 with the Fibonacci series of incredibly large number of terms (in practice, not only for mathematical infinity).
Therefore, an error would be a chance for a number of terms that might occur in practice. The increased frequency of the CPU and consequently reduces its precision errors occured when searching for the next term of the series.
But errors can occur in other programs due to loss of precision processor or make the system unstable.
Something similar happened to me. The Prime95 gave error with the E2180 to 3GHz (cooler box), with vcore of 1.31. I thought: "if I raise the vcore too, this man is going to roast hot." Was already hitting 78 degrees in Prime95. FPu was in error if I am not mistaken. Then decided to test the stability with the folding SMPS, using 100% of the nucleus. It was 10 hours and not had any error. The temperature was hovering between 62 and 66 degrees. My cabinet is poorly ventilated. Also test matches and stayed for good. Finally, with the Prime 95 and left the proc to 3GHz same.