General Questions About How to Preform Benchmarks
#1
So out of my own curiosity I want to see how much of a performance boost AMD's Phenom II processors get from overclocking since most of the benchmarks here on the forums are from intel cpus. So I have a couple of questions before I start on this journey.

1. Which of the games that I have will make good games to benchmark? I currently own:

Final Fantasy X
Final Fantasy XII
Dragon Quest VIII
Odin Sphere
Xenosaga Episode I
Virtua Fighter 4: Evolution
.hack// Part 1
Burnout 3
Kingdom Hearts
Kingdom Hearts 2
Kingdom Hearts Re: Chain of Memories
Shadow of the Colossus
Star Ocean 3 (in the mail from Amazon and I will not get it for a week+ so I would have to add this one later)
Suikoden 3 (in same boat as Star Ocean 3)

2. Is there someway for Fraps to measure the average FPS while I am playing a game? I'm sure its possible I just don't know how to do it.

3. I'm assuming the correct way to do this is to not have any speed hacks on while playing the game, right?

In the end I want these benchmarks to help fellow AMD buyers see how overclocking would help them or if its not worth it after you pass a certain speed (ie you don't get much better performance between 3.2GHz and 3.4Ghz). Also, I want to see if its worth getting a top of the line air cooler in order to push my system past 3.4GHz. Here is the system I will be using:

AMD Phenom X3 720 OC @ 3.4GHz (I might be able to push it to 3.6GHz just for testing purposes since PCSX2 only uses 2 cores so the temps are alot cooler while using it then stress testing with Prime 95)
ATI Radeon 4870 1GB version (might attempt some benchmarks with this overclocked to see if it actually helps to overclock the GPU)
4GBs of DDR3 RAM
Windows 7 RC

Any suggestions of how I should do these are welcomed (also if somebody recently did this but I couldn't find it through searching please point me to the thread).
Reply

Sponsored links

#2
Any or all of those games can be used as a benchmark, just remember to turn Frame limit off.

Good benchmarking practice assumes that you do each test at the EXACT same spot with an easily repeatable setup. Find a static (IE not a lot happening or has a lot of variables like people randomly showing in and out of screen). Safest way to do this is to make a savestate of the spot you want to test.
[Image: 2748844.png]
Reply




Users browsing this thread: 1 Guest(s)