Does hard drive matter in benching?

[luihed]

Will an old 20g 5400rpm hard drive affect my benching results?.... Im going to do some benching soon and thats the only hard drive I can use to bench..... Will bench Superpi and 3dmark only... thanks


lui

[theelectic]

Not for those benches, no.

[xlink]

yes, but it's almost negligible. for SP1m it pretty much wont' affect it whatsoever.

[Nanometer]

yes, but it's almost negligible. for SP1m it pretty much wont' affect it whatsoever.

SP1 does not use HDD access during the benchmark. If yours is then you must be either using 64megs of ram or doing something seriously wrong.

[Xope_Poquar]

The slower your hard driver the more time it takes to load each benchmark (for 3dMark). That means you have longer for your video card to cool down.

[cadaveca]

Makes big difference in pcmarks, small difference in 3dmarks, and no difference in PI.

[jo2007]

Im kinda new but i want to ask for benchers do they use the same hard drive so they can load their softwares ( 3dmark, super pi, etc ) so when they boot windows will load to automatic desktop with softwares im lost at that point thanks XS.

[INJViper]

I do most if not all of my benching on a small HDD running XSOS v.1 that way there are the least amount of programs and crap running in the background.

[xlink]

SP1 does not use HDD access during the benchmark. If yours is then you must be either using 64megs of ram or doing something seriously wrong.

it inevitably does in some way or another.

note the word negligible, as in insignificant. chances are it wouldn't even show up to the millisecond.


also, what's wrong with running 3 instances of memtest with SP simultaneously?

[dinos22]

if you had a hardware hard drive controller on a PCI card >> that would make more difference in performance for SPI as it reduced CPU overhead otherwise shouldn't be much

but it would be interesting to compare 5400vs7200 drives

[wwwww]

SP1 does not use HDD access during the benchmark. If yours is then you must be either using 64megs of ram or doing something seriously wrong.

Yes it does, after it finishes the calculation, it writes the result time and then gives your score.

I'll do a sample shortly.



SuperPi on a network drive. Don't ask me why it took so long to write a 1MB file, the point is that there is a difference, if your drive is slow enough then you have a major difference.

[Vapor]

It DEFINITELY uses the HDD. See all those .dat files it writes during the run?

Those are actively written and read from during the entire benchmark. This is why ramdrives are faster and why network drives (and old flash drives) are just POKEY.

I don't have "show file extensions" enabled on this system, but here they are:

[Nanometer]

well i guess it does matter. I figured the files were written to the ram..

[DFI pit bull]

Interesting thread, this is something that I have been wondering for a long time.
Has anyone tested this theory? i.e done benchmarks with various drive types and setups, would be keen to know.

[apexracing]

Perhaps it dumps to HDD after the test is done, but no HDD during the test.

[jimmyz]

try it from a usb drive if you have a while.lol it is super slow.

[lowfat]

SP1 does not use HDD access during the benchmark. If yours is then you must be either using 64megs of ram or doing something seriously wrong.

yes it sure does affect it. Have you ever ran SuperPi while installed on a ram drive? It definately does help. It helped a good half second with my ol' Pentium M 750. Haven't tried with anything since, but it sure use to help.

[GoldenEye]

If you run SuperPI on a iPod or on a intern HDD SATA @ 7200 RPM it make a different