how do passmark comprise the charts?

[Guj]

Hiya guys, long time reader, first time poster

Can somebody explain something to me please. what happenes if 2 people from across the world on identical computers run the performance benchmark and upload it?

person A has a clean install of windows, drivers, directx, passmark etc.
however person B has a totally fragmented HDD, running the test in the baking desert with the virus scan running in the background?

how are the scores added to the charts?

Regards, Guj

[David (PassMark)]

Some results are filtered when they are submitted. Examples are when the submission does not contain a full set of test results, when a CPU os overclocked, or when the submission is from the same PC as a result already in the database.

Otherwise, for the CPU, Disk and video charts, if a result doesn't get filtered, it gets averaged with all the other results of the same type.

So for example you can get CPU results being pulled up slightly by using 64bit O/S and good RAM, or being pulled down by background tasks running on 32bit.

But the averaging process means you don't see the extremes in general. For the popular CPUs we typically have between 1000 and 10,000 different systems being averaged to get a single result for that CPU type.

[2006]

You don't display the megahertz anymore though.

[David (PassMark)]

With the turbo mode features in AMD and Intel, and associated the power saving under clocking options, the average Mhz speed was starting to look silly. People didn't understand that the Mhz speed for new CPUs continuously jumps about in a large range. We got a lot of EMail telling us the Mhz speed numbers where wrong. They weren't wrong, but they weren't all that useful either.

[NotMeAgain]

are different versions of passmark taken into account, too?

I do not understand, why some older drives get better ratings then newer Drives with better technology and features.

My example, I am searching a 120GB IDE HDD with 7200rpm:

WDC WD1200LB - passmark 410 (three platter design, 2MB Buffer)
WDC WD1200SB - p 389
Hitachi HDS722512VLAT80 - p 352
WDC WD1200PB - p 309
WDC WD1200JB - p 292
Hitachi HDS721612PLAT80 - p 278 (two platters, 8MB Buffer)

[David (PassMark)]

The results are all from V7 of PerformanceTest.

We tried to create a test that isn't greatly effected by caching.

Having 3 or 2 platters has more impact on the capacity than the speed.