New Dolphin CPU Benchmark - NO GAME REQUIRED

[Erus_Iluvatar]

CPU: Intel QX9300M @ 2.53GHz
OS: Windows 7 x64
Time: 22 minutes, 11 seconds
Screenshot: http://i.imgur.com/A2hfuIN.png

[Nintendo Maniac 64]

(01-18-2014, 09:31 PM)Ian Cutress Wrote: As you can imagine, NM64 emailed some review websites about this benchmark.

Shhhhh, I'm not affiliated with the benchmark nor the Dolphin devs at all. For all I know the devs may not like the extra publicity...

[Anti-Ultimate]

(01-18-2014, 09:26 PM)delroth Wrote: Gahh, OC is so fucking complicated to take into account that I'm wondering if I should even include in there. It screws up everything if people overclock cache and some not for example.

Just add a collum for "Has overclocked Cache Ration?" for Sandy -, Ivy Bridge and Haswell cpu's. I really want my result included in there :/

[delroth]

(01-18-2014, 09:39 PM)Nintendo Maniac 64 Wrote:

(01-18-2014, 09:31 PM)Ian Cutress Wrote: As you can imagine, NM64 emailed some review websites about this benchmark.

Shhhhh, I'm not affiliated with the benchmark nor the Dolphin devs at all. For all I know the devs may not like the extra publicity...

I don't really care as long as people understand what this benchmark is about: it's very single core performance oriented (you're unlikely to see any difference from multicore here) and tends to kill branch prediction in most CPUs due to it running inside a JIT with a large amount of generated code. Not surprised to hear that RAM/Cache speed makes a huge difference as well: with the whole block behavior of the Dolphin JIT there are a huge number of loads/stores to ppcState, which should fit in L2 on most CPUs.

[Anti-Ultimate]

I forgot to include what my Cache Ratio was overclocked to >.> It's at 40 right now but I plan to push it higher.

[Nintendo Maniac 64]

Someone totally needs to try running this on a Wii U and see if there's any difference between it and a normal Wii.

[delroth]

(01-18-2014, 09:47 PM)Nintendo Maniac 64 Wrote: Someone totally needs to try running this on a Wii U.

vWii runs at the same clock speed as the original Wii, on a CPU of the same architecture. I don't expect any difference. Would be interesting if there is.

FWIW, 2 different people have tested on their Wii and got the exact same time.

[Anti-Ultimate]

(01-18-2014, 09:47 PM)Nintendo Maniac 64 Wrote: Someone totally needs to try running this on a Wii U and see if there's any difference between the vWii and a real Wii.

I'm on it. (If I manage to get it onto my SD card)

EDIT: Turns out my WiiU doesn't like SD cards anymore and wants them formatted. But to be honest, I'm fairly sure there isn't going to be any difference.

[delroth]

(01-18-2014, 09:49 PM)Anti-Ultimate Wrote:

(01-18-2014, 09:47 PM)Nintendo Maniac 64 Wrote: Someone totally needs to try running this on a Wii U and see if there's any difference between the vWii and a real Wii.

I'm on it. (If I manage to get it onto my SD card)

Use http://delroth.net/povray.elf on a real console (fixes a bug in the shipped .elf which isn't noticeable on Dolphin).

[Anti-Ultimate]

I for the Lord of god can't get this thing to be recognized in HBC. What am I supposed to do^^