Hitorical Base lines and outliers

[hespozel]

I Notice that some CPU scores are influenced by "outliers", where very low baselines scores take the average down. The results are that CPUs that are better than others have lower average. Is it true ? for example i7-8665U has a lower CPU Mark average than a i7-8565U. This not make sense !. Another fact. The i7-8665U a month ago, as an average score much higher than now. Is there a way to fix that ? Is there way to see all CPU baseline of a specific processor ?

Thanks. Heraldo

[David (PassMark)]

Those two CPUs are so close that other factors move the result around a lot more than the tiny difference in clock speed. (e.g. RAM used & how aggressive the power saving settings are)

You can see the distribution in PerformanceTest.
All CPUs have some units at the high and low end. Especially laptops where power control, thermal issues and bad RAM configs are common.

These CPUs aren't special in that regard.

[hespozel]

I still state that the CPU average for these two processors are not reflecting that i7-8665U have a better performance than i7-8565U. My suggestion is to remove scores that are outliers, to have a average score that reflects the correct processor speed. It is easier to overload a machine with lots of tasks running and have very poor results than set the OS environment to have good results. For example for both processors, does not make sense to have scores less than 8000 points ...

[David (PassMark)]

There are outliers for all CPUs. They average out in the end once there is enough samples.

Again, the on paper performance difference for these CPUs is a small 5%. And a good or bad RAM setup, or having a thermal envelope set in BIOS can make a much larger difference than this. The i7-8565U will generate a bit more heat and getting rid of heat is a big issue in laptops.

What it means is that you will not (on average) benefit from upgrading to a i7-8565U CPU and we think it is important information to make available. Not something we need to hide.