[Unofficial] New Dolphin 5.0 CPU benchmark - results automatically updated!

[Nintendo Maniac 64]

(07-06-2016, 03:00 PM)Bluescreendeath Wrote: Thanks. I just downloaded from your link and ran the povray benchmark at 2.0GHz and 2.4GHz.

But did you run it in Dolphin 5.0?

UPDATE: Well I just tried running povray in Dolphin 5.0 and it simply does not seem to work. I've contacted delroth to see if he knows anything about this...

In the mean time, I've added 5 more benchmark results - one from a T5800 Merom and four from an i5 520M Westmere. So far it looks like my L2 cache theory may be correct.

[Bluescreendeath]

(07-07-2016, 11:54 AM)Nintendo Maniac 64 Wrote: But did you run it in Dolphin 5.0?

UPDATE: Well I just tried running povray in Dolphin 5.0 and it simply does not seem to work.  I've contacted delroth to see if he knows anything about this...

In the mean time, I've added 5 more benchmark results - one from a T5800 Merom and four from an i5 520M Westmere.  So far it looks like my L2 cache theory may be correct.

I think I got it working. I copied and pasted the Dolphin 5.0 files into the POV benchmark folder that had the older Dolphin version. When I started the program, it said "Dolphin 5.0" - and was able to press the play button to run the benchmark (pic attached).

These are the results I got.

i7 2600k @ 2.0GHz - 16min 37 sec (997 seconds)
i7 2600k @ 2.4GHz - 15min 52 sec (952 seconds)


Do you think the L2 cache problem is just a problem with this benchmark, or does it translate into issues with Dolphin 5.0 as a whole?

[Nintendo Maniac 64]

Turns out that povray needs an SD card which is why it wasn't working.


Unfortunately, there's not much guarantee that we're using the same Dolphin configuration, so I've made a quick modification to the Dolphin 5.0 CPU benchmark bundle to also include the "old povray bench":
https://drive.google.com/file/d/0B3hUE5vj6-14Ukt1d0JPdWRBSlk

Please try benching with that, and sorry to keep using you as a guinea pig. You don't need to run the luabench though because it'll perform exactly the same as the configuration in the first post (I tested it already).



Now as for my own results, I only benched the Penryn and the Westmere at 2.0GHz due to time constraints.

C2D T8300 Penryn - 17m 06s (1026 seconds)
i5 520M Westmere - 15m 17s (917 seconds)

Needless to say, I think we're really going to have to break out something like the old Wind Waker benchmark to see which (if any) of these two benchmarks actually correlate to Dolphin's real-world relative performance.

[Bluescreendeath]

I downloaded your file and ran the default (luabench?).

2600k @ 2.0GHz = 18min 4s (1084s)
2600k @ 2.4GHz = 14min 57s (897s)

[Nintendo Maniac 64]

(07-08-2016, 03:32 PM)Bluescreendeath Wrote: I downloaded your file and ran the default (luabench?).

No...that's the wrong thing - run the povray benchmark.  You can run it by double clicking "Run old povray benchmark".

Also, if you hit "Stop" in Dolphin, you'll see povray.elf listed in the game list and you can select it from there.

And you really don't need to run at 2.4GHz when I didn't even run povray at that.

[Bluescreendeath]

Thanks, I found that out right after I posted. Here are the benchmarks with the old pov.

Times with old pov benchmark:

2600k@ 2.0GHz = 13min 43s (823s)
2600k@ 2.4GHz = 11min 30s (690s)


Thank god I'm using software overclocking...otherwise, having to reboot to BIOS every time I wanted to change my clock speed would've driven me crazy.

[Nintendo Maniac 64]

Yeah OK, we definitely need to compare with a real game benchmark or something, because the delta between Penryn and Sandy Bridge in povray and luabench is ridiculously different.

Do you happen to have any of the following games?

Legend of Zelda: Twilight Princess (Wii)
Legend of Zelda: Wind Waker
New Super Mario Bros Wii
Nintendo Puzzle Collection
Star Wars Rogue Leader
Super Mario All Stars 25th Anniversary
Super Mario Sunshine
Super Smash Bros. Melee
Wii Sports

[Bluescreendeath]

(07-08-2016, 04:15 PM)Nintendo Maniac 64 Wrote: Yeah OK, we definitely need to compare with a real game benchmark or something, because the delta between Penryn and Sandy Bridge in povray and luabench is ridiculously different.

Do you happen to have any of the following games?

Legend of Zelda: Twilight Princess (Wii)
Legend of Zelda: Wind Waker
New Super Mario Bros Wii
Nintendo Puzzle Collection
Star Wars Rogue Leader
Super Mario All Stars 25th Anniversary
Super Mario Sunshine
Super Smash Bros. Melee
Wii Sports

I have SSBM and I think I have Zelda TP. How would I benchmark these games?

[Nintendo Maniac 64]

Hmmm... this is going to be more difficult than I though.

Turns out, the Intel integrated graphics on the Penryn machine is so el-lame-o that 1. it can't even launch SSBM and 2. idling on the title screen for Nintendo Puzzle Collection (a 2D game!) runs at ~15fps while I get ~75fps on the Westmere (which has a discrete GPU). Yes, both CPUs were running at 2GHz.

Last time I checked, a Core 2 Duo isn't 5x slower than a Westmere in anything.

[PortlandStandard]

Can you update the sort? New high score, did I break it?