Dolphin, the GameCube and Wii emulator - Forums

if you really want some true performance metrics, you should run a fifo log in the same area across many different computers with different hardware all set to defaults.

(07-08-2016, 01:14 AM)KHg8m3r Wrote: [ -> ]if you really want some true performance metrics, you should run a fifo log in the same area across many different computers with different hardware all set to defaults.

That's kind of what I was thinking with regards to running the old Wind Waker benchmark in Dolphin 5.0, but what you suggest would indeed be better.  I'm not a developer or tester or anything so it never crossed my mind to look into using such tools for benchmarking purposes.

(07-07-2016, 03:17 PM)Nintendo Maniac 64 Wrote: [ -> ]Even if all Core 2 CPUs were the slowest thing on Earth, a Core 2 Duo Wolfdale would still be faster for Dolphin and would overclock better than a Core 2 Quad.

Besides, a large reason why I mentioned it is because said CPUs are only around 15 bucks with shipping anyway - I imagine cost would be very important because otherwise OP would likely not be using the PC parts that they are currently using.  If such a CPU cost 50 bucks then yeah screw that.


Regardless, it's not just the 2600k showing slower performance - there's also an i5 520M (Westmere) and an i3 3110M (Ivy Bridge).  For reference, I plan to do more performance comparisons in coming days and weeks with the previous benchmarks (povray and Wind Waker) in Dolphin 5.0 to see if this behavior is in fact consistent.

Also, I'm not sure I trust that 5500U (Broadwell) result when the Haswell results are much faster per clock...that's arguably even weirder than Core 2 being faster than Ivy Bridge.  But unlike the Core 2 results, there's not really any reason for a Haswell desktop owner to upgrade to Broadwell when the 4790k exists, so that performance discrepancy shouldn't really be of any real-world concern anyway.

I am certainly unhappy with this situation. I hope it's only because of the benchmark and not reflective of Dolphin 5.0 in general.

I really think that it is just this test that is skewed towards a large L2 cache.
You don't need the WInd Waker game to run the Fifilog (unless something about copyright)
But then the question is, how long do you make the log, and how to filter out the effects of GPUs

(07-08-2016, 12:12 PM)KHg8m3r Wrote: [ -> ]You don't need the WInd Waker game to run the Fifilog (unless something about copyright)

Regarding Wind Waker, I was referring to the old old Dolphin benchmark:
http://forums.dolphin-emu.org/Thread-the-legend-of-zelda-the-wind-waker-cpu-benchmark

It's as bad as your B940 . The iGPU is also below the minimum GPU requirement
If you can disassemble your laptop then i5 2520M @ 2.5->3.2GHz max turbo($26) should be a much better choice for the same socket