Bug in CPU frequency calculation.
Posted: Thu Jan 15, 2015 12:34 pm
Hi.
I try to overclock my old and still using amd 64 3500+, stock @ 200 * 11 = 2.2Ghz.
So if i keep 11 multiplier, the frequency show in Core Temp are correct.
But if i use 10.5 multiplier the frequency differ from expected and what CPUID show :
What i suspect is the calculation is wrong because not only the multitiplier is reduced from stock value (like expected) but also the FSB is accordingly reduced by the same ratio ! So the final frequency is too much reduced actually.
In my case : the multiplier is reduced by : 10.5*100/11 = 95.45% and in Core Temp the FSB is also reduced by this ratio : 240*0.9545 = 229.1 !!!
So Bug i suppose.
I try to overclock my old and still using amd 64 3500+, stock @ 200 * 11 = 2.2Ghz.
So if i keep 11 multiplier, the frequency show in Core Temp are correct.
But if i use 10.5 multiplier the frequency differ from expected and what CPUID show :
What i suspect is the calculation is wrong because not only the multitiplier is reduced from stock value (like expected) but also the FSB is accordingly reduced by the same ratio ! So the final frequency is too much reduced actually.
In my case : the multiplier is reduced by : 10.5*100/11 = 95.45% and in Core Temp the FSB is also reduced by this ratio : 240*0.9545 = 229.1 !!!
So Bug i suppose.