• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

CPU benchmark programs?

V

verymagicalguy

Limp Gawd
Joined
Mar 28, 2008
Messages
204
I'm trying to analyze the affect of cache and its affect on the speed of different type of processing. I know that these types of reviews have been written before, but I wanted to see for myself (and possible share with you guys) what the difference in cache actually has on real world performance. In my possession I have a E2160 (1mb), E4500 (2mb), and E6600 (4mb). I will clock all of them to 9x300 = 2700mhz and test the affect of the different caches on performance.

What programs/types of tests should I run? Also, is there a reliable way to test things like loading programs? Is there an effective way of keeping type besides with a stop watch? It would seen difficult given the speed these applications would load.

So far the only one I've found is a video encoding one. I could also do a WinRAR extraction as well as SuperPI. Are there any special programs that will run tests? What else should I do for an all inclusive review? Thanks!
 
SuperPI 1M and 32M tests, CPU scores from 3dmark06, Cinebench, CPUmark from ZDnet, SiSoftware Sandra...
 
Personally, I tell people to avoid SuperPi like the plague: it's single threaded, not multicore aware, and basically old as the dickens. The new Sheriff on the "prime" testing block is wPrime, properly multithreaded and getting more popular all the time.

Testing one core of a multicore processor with SuperPi to me is like trying to make a Ferrari work with 1 wheel attached. It just doesn't do the job quite right.

As for something that will test only the CPU and RAM, one of the best and most simple benchmarks I've ever come across that has very repeatable results is Geekbench. The creators now charge for a newer "improved" version of the software, but the older version - tough to come by anymore - was free and so I'm posting it HERE for anyone that wants it.

Virus checked with NOD32 v3 (fully updated 20 mins ago) before I archived it into a Zip file, so it's clean. There's a geekbench32 (32 bit) and a geekbench64 (64 bit) command line only console application inside the archive. Open a Command Prompt and run them in sequence, see what you get. I've been using those two little apps for years now, with great results.

3DMark is still primarily a GPU test (and a god damned huge one considering the 650MB+ download, insane) but does offer some CPU benchmarking info when the testing is done. Cinebench is another nice one, but still primarily a GPU/video card test. Sandra, Everest, and PCWizard all have CPU benchmarking routines in them, also.

I'd say use Geekbench on each processor, in 32 bit and 64 bit versions, and go from there. They run fast, give detailed info about specific processing speed (FLOPS info for CPU, RAM bandwidth info for RAM) and in general they just kick ass. :)

Good luck...
 
He wants to see if the amount of L2 affects speed. Dual core or not, SuperPI would be a nice test either way. Number of cores and the clock speed it the same, only variable changing is the amount of L2... SuperPI would be a great candidate. But if you want to see raw CPU power, Sandra.
 
You must log in or register to reply here.
Back
Top