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

[Nintendo Maniac 64]

(07-03-2016, 03:46 AM)Bluescreendeath Wrote: I think I screwed up somewhere. Did I download the wrong benchmark? My final results screen seems to be different from that of other people's.

That to me looks like the old Povray benchmark.  Did you try the benchmark + Dolphin bundle for Windows users rather than the manual method?

Also, make sure the folder you went into was named "Dolphin 5.0 CPU Benchmark".

(I will likely add screenshots of these things to the first post)

And for your convenience, I've also removed your two benchmark scores since they're clearly not correct...

[Bluescreendeath]

(07-03-2016, 06:21 AM)Nintendo Maniac 64 Wrote: That to me looks like the old Povray benchmark.  Did you try the benchmark + Dolphin bundle for Windows users rather than the manual method?

Also, make sure the folder you went into was named "Dolphin 5.0 CPU Benchmark".

(I will likely add screenshots of these things to the first post)

And for your convenience, I've also removed your two benchmark scores since they're clearly not correct...

Thanks, I redownloaded the benchmark and it works fine now. Just uploaded a G3258 @ 4.0GHz benchmark.

The scaling seems extremely linear in comparison to your 2.0GHz G3258 benchmark. I got 378 seconds for 4.0GHz - you had 755 for 2 Ghz. 755/378 = 1.99
100% faster Ghz translating directly into 100% faster benchmark speeds.

[Bluescreendeath]

If anyone here has Sandy Bridge or Ivy Bridge CPUs, I'd appreciate it if they could submit benchmarks for them at various speeds. My Sandy Bridge i7-2600k is running very slowly (compared to the Core 2s) with this new Dolphin Benchmark. 8-9 minutes at 4.0GHz, and around 18 minutes at 2.0GHz.

I've ran it multiple times at 2.0GHz, and always gets around 18 minutes - worse than the IPC of a Core2. I ran passmark and my cpu got 8670.4 - which is around the average score for the i7 2600k on passmark. So it seems my cpu is running fine in other programs. Is it possible that Dolphin 5.0 nerfed Sandybridge performance and buffed Core2 performance?

[KHg8m3r]

(07-04-2016, 04:03 PM)Bluescreendeath Wrote: Is it possible that Dolphin 5.0 nerfed Sandybridge performance and buffed Core2 performance?

Dolphin doesn't know what CPU you have, so that's impossible
What's more likely is that your computer is doing something weird. Make sure your windows power management is set to high performance

[Nintendo Maniac 64]

(07-05-2016, 05:38 AM)KHg8m3r Wrote: Dolphin doesn't know what CPU you have, so that's impossible

I don't think he means intentionally nerfed so much as the new Dolphin for whatever reason likes the Penryn architecture vs Sandy Bridge - maybe it's the larger L2 cache, or the lack of SMT?

[leolam]

Technically, Dolphin does know what CPU you have (how could it display it in the OSD and send it for analytics otherwise?)
But of course it's not doing anything intentionally to slow down specific CPU generations.

[KHg8m3r]

(07-05-2016, 06:51 AM)leolam Wrote: Technically, Dolphin does know what CPU you have (how could it display it in the OSD and send it for analytics otherwise?)
But of course it's not doing anything intentionally to slow down specific CPU generations.

Analytic yes, OSD, no, as that is only the GPU

[Bluescreendeath]

(07-05-2016, 05:38 AM)KHg8m3r Wrote: Dolphin doesn't know what CPU you have, so that's impossible
What's more likely is that your computer is doing something weird. Make sure your windows power management is set to high performance

(07-05-2016, 05:52 AM)Nintendo Maniac 64 Wrote: I don't think he means intentionally nerfed so much as the new Dolphin for whatever reason likes the Penryn architecture vs Sandy Bridge - maybe it's the larger L2 cache, or the lack of SMT?

As Nintendo Maniac mentioned, what I meant was Dolphin 5.0 liking the Core2 architecture (eg. larger L2 cache, lack of hyper threading, whatever) so it runs more efficiently on certain Core2s than my Sandybridge i7.

[Nintendo Maniac 64]

One thing that may be of interest to try out is to in fact run the old Povray.elf homebrew benchmark but inside the preconfigured Dolphin 5.0 bundle that I've provided in the first post; this way we can make sure that it is Dolphin 5.0 and not the luabench that's at fault.

If you do this, then please make sure that you configure your Sandy Bridge to have a clockrate of no higher than 2.4GHz so I can run it on my Core 2 Duo for direct comparison.


If you need a copy of the povray benchmark, you can get the old Dolphin povray benchmark bundle here; the file you want is called "povray.elf": http://delroth.net/Dolphin-Benchmark.7z

[Bluescreendeath]

(07-05-2016, 05:27 PM)Nintendo Maniac 64 Wrote: One thing that may be of interest to try out is to in fact run the old Povray.elf homebrew benchmark but inside the preconfigured Dolphin 5.0 bundle that I've provided in the first post; this way we can make sure that it is Dolphin 5.0 and not the luabench that's at fault.

If you do this, then please make sure that you configure your Sandy Bridge to have a clockrate of no higher than 2.4GHz so I can run it on my Core 2 Duo for direct comparison.


If you need a copy of the povray benchmark, you can get the old Dolphin povray benchmark bundle here; the file you want is called "povray.elf": http://delroth.net/Dolphin-Benchmark.7z

Thanks. I just downloaded from your link and ran the povray benchmark at 2.0GHz and 2.4GHz.

i7 2600k @ 2.0GHz = 22 minutes 49 seconds (1369 seconds)

i7 2600k @ 2.4GHz = 18 minutes 46 seconds (1126 seconds)

Pics attached.