Dolphin, the GameCube and Wii emulator - Forums

Alright, I'm done with the 64-bit benchmark using Linux. Honestly, that was...an unpleasant experience, to say the least.

All in all, I would qualify the benchmark as worthless. I couldn't figure out how to install Dolphin 3.5 and attempting to access the OP's 'Use these settings' page gives me a restricted message. So basically, the entire test was completely compromised. Regardless, the results weren't that different anyway. Save #1 was virtually the same, save #2 got a noticeable increase, and save #3 got a noticeable decrease -- and these differences can probably be attributed to the differences between Linux/Windows and 4.0/3.5 anyway.

For what it's worth, here are the results for fun.

OS: Linux Mint 15 1.8.6 64-bit
Game Region: NTSC
Emulator: Dolphin 4.0-263
CPU: AMD Phenom X3 8650, Triple Core, 2.30Ghz (Stock)
GPU: Nvidia 9800GT, 512MB GDDR3 (Nouveau driver 1.0.7; trying to use Nvidia's proprietary drivers would cause a system crash)
RAM: Micron 2x1GB DDR2 PC2-5300

Results (FPS/VPS/Speed):
Save #1: 26/52/87
Save #2: 27/53/89
Save #3: 21/43/72

And here are the images:
Image 01
Image 02
Image 03

---

(10-22-2013, 09:27 PM)Jhonn Wrote: [ -> ]MAke sure you run on AC power, my laptop, for example, won't allow my CPU running higher than 2.8GHz on battery, even with "high performance" settings...

Good news. I tried it and it turns out you're correct. I ended up getting a few extra fps while keeping it plugged in.

Now the bad news -- I discovered that I can't change the power settings, which must have been what was happening before and I hadn't noticed because I just presumed that it had saved on my command. I researched it and, apparently, this appears to be a quirk with Asus laptops (particularly Windows 8 ones) in which there is some background program that is overriding user commands and automatically switching settings in order to ensure battery/performance as needed. As it is, I can't seem to figure out which program is doing it and how to disable it, so I'm going to edit my Asus benchmark and list it as compromised.

So all in all, three tests later, it seems I'm left with only one usable benchmark.

in this test only dolphin version 3? or 4.01 also?

Different versions have different performance, so using 4.01 will give different results than everyone else.

Also it seems a bit silly to re-test on 4.01 when...
1. we already have quite a few results from version 3
2. a benchmark mode is being considered to be added to new revisions (LINK)

I know I'm late, but:

OS: Windows 7 64-bit

Game Region: NTSC

Emulator: Dolphin 3.0 64-bit

CPU: I5 2300 @ 2.9 ghz (Stock)

GPU: MSI GTX 560

RAM: Corsair Vengeance 8GB 1333 mhz

Save #1

FPS/VPS/Speed percentage

58/117/195%



Save #2


FPS/VPS/Speed percentage

47/96/160%


Save #3


FPS/VPS/Speed percentage

74/149/248%

Well, I decided to join this thread. but I noticed that the emulation speed drops a LOT after a few seconds.
i upload screenshots after the fps drop, ok? (I can upload screenshots without the fps drop, too)

Game region: NTSC-U

speed before FPS drop: 180FPS, 360 VPS, 600%
after FPS drop: 73FPS, 146VPS, 243%



speed before FPS drop: 201FPS, 403 VPS, 671%
after FPS drop: 59FPS, 118VPS, 196%




speed before FPS drop: 169FPS, 338 VPS, 563%
after FPS drop: 92FPS, 184VPS, 306%




specs:
CPU: Intel Xeon e5 2687w | 8 cores/16 threads @ 3.1GHz (3.8GHz turbo)
GPU: GeForce GTX 260 by Zotac
6GB RAM (Temporally)
OS: Windows 7 x64
Dolphin is saved on an ssd by OCZ (Vertex 3)

For some people, I think it's normal that the speed drops after the first few seconds. For the benchmark, just wait until the speeds stabilize and hold steady (more or less). It seems like you've done that by recording after the drop, so I don't see any problems with your results. Thanks for contributing, btw, not sure when Garteal will get the chance to update the spreadsheet with all the results though.

My system: AMD Phenom 9550 @ 2.2 GHz, 6 GB generic DDR2, GTX 660, win7 x64
Test 1: FPS: 26 - VPS: 54 - SPEED: 90%

Test 2: FPS: 23 - VPS: 47 - SPEED: 78%

Test 3: FPS: 34 - VPS: 68 - SPEED: 113%

Can someone post the result of a sandy/ivy bridge and a haswell at the same clock speeds?

(01-01-2014, 04:18 PM)drhycodan Wrote: [ -> ]Can someone post the result of a sandy/ivy bridge and a haswell at the same clock speeds?

The chart linked in the first post already contains that. In particular, note the i5 2400 and i5 2500K @ 3.4 GHz, the i5 3570 @ 3.4GHz, and the i5 4670 @ 3.4GHz.

You can also use basic algebra to determine performance at a given clock rate based on performance at another clock rate. This works since performance has been shown to scale linearly with clock rate in this benchmark.