Dolphin, the GameCube and Wii emulator - Forums

Full Version: CPU thread
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
Can't really contribute a lot myself, so anyone feel free to hijack the thread or make another one. Basically, I thought we'd give a short review of most popular CPUs. OC data could probably help too. (I don't have). This is assuming use of the latest revision, and supposedly the "best" settings (no game-breaking shortcomings, such as missing sound with HLE)

Intel E8400 @3.00GHz: (tested in win7 x64)
-Will run many lesser demanding games fullspeed, like Pikmin 1/2, Super Paper Mario, Super Smash Bros Melee, Super Mario Sunshine, Zelda: Wind Waker, Mario Kart: Double Dash, Muramasa
-Will run many games consistently at fullspeed or very playable framerates, like Paper Mario 2/Super Smash Bros. Brawl
-Suffers huge slowdowns in demanding games, like Metroid Prime 1/2/3, Super Mario Galaxy 1/2, Zelda: Twilight Princess, Metroid: Other M, Okami

Alternatie suggestion: Make a requirement scale for all games, and just rate the CPUs with a strength. (I know there's a wiki, but I don't think many people use it)
A few suggestions for you if you want to have usable/reliable data if I may:
1. Have them all use the same build, provided by you.

2. Provide a separate 32 bit and 64 bit build and have them post which build they used.

3. Have them all use the same standard settings, adjust the settings to be more cpu heavy and less gpu heavy to reduce chances of interference from outside bottlenecks.

4. Make sure they include what OS they are using.

5. Make sure they provide memory, cpu, and gpu specs so that you can go back and check for bottlenecks if some of the results don't seem right.

6. Have them all use a standard game and a specific spot from a savegame file.

7. Have them list the vps and gamespeed in windowed mode with framelimit off.

If enough people where to posts their results under these rules you could compile some very nice graphs to study the relationship between performance and different architecture and variables (it worked very well with the GPU thread before a link to the past went insane and started deleting everything).

I was actually hoping someone would get around to making a thread like this.
Pretty much everything that NV said.
One thing to add though, we should have a (small), from demanding to easily emulated games that everyone should use to test the CPU, where we all use the same settings for each game.

You should update the first post with the newly added data if you can.