Different results for the same CPU

[Abbas]

I'm testing three netbooks- the Toshina NB200, the HP Mini 110 Studio tord and the Lenovo ideapad S10-2. All of them have the Intel atom N280 CPU running at 1.67GHz

Running the CPU test, the Toshiba scores a 350.90, the Lenovo scores 320.90 and the HP scores a 219.50. Why would there be such a difference in CPU scores when all three netbooks are running an identical CPU.

This makes me prevent using your app for publishing results.

[David (PassMark)]

For increased accuracy you need to repeat the test several times and discard low results, which may be caused by background tasks running, etc..

But it is naive to expect different machine from different manufacturers to have identical score.

BIOS differences, motherboard differences, memory differences, high temperatures, different CPU steppings, and other software running on the machine can all have an impact on the CPU results. You can blame the test is you want, but you might be better served looking for the real reason the performance is different.

[Abbas]

Thanks- but what would "several times" constitute of? Will 5 times be enough to get a reasonable score?

I can understand if the overall performance differs from machine to machine, but I would think that a CPU test should similar results (<3% difference) across different machines in that CPU is identical.

I can understand that background tasks could come in the way of performance though.

[David (PassMark)]

Yes, 5 should be enough. You should find the results are simailar for each run, unless there are some external factors (e.g. CPU throttling due to heat build up, or background software tasks running from time to time).

Another thing to look out for in laptops is the O/S power management settings. The laptop might be configured to save power, rather than for max performance. This is fairly common and can effect CPU performance.