Dolphin, the GameCube and Wii emulator - Forums

Full Version: Why doesn't any hardware review sites use Dolphin to benchmark new CPUs?
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
Pages: 1 2 3
Seriously, Dolphin is not only an excellent emulator that sets the milestone to be the first working emulator for a 7th gen console, but it also makes a great benchmark tool to measure a CPU's IPC. It is one of the best tools I've used that can stress the heck out of a CPU based on it's IPC and single threaded performance. After all, it is the CPU's IPC that counts most when it comes to measuring a CPU's raw power.
Because none of them have ever heard about Dolphin? Now that you mention it, I would suggest creating a new website section called " Dolphin-based CPU ranks" or something like that; and them put there the latest benchmark done (delroth´s).
Haswell is 30% faster than Ivy Bridge in Dolphin . Does that ring a bell ?
Haswell isn't that much faster than Ivy Bridge in most benchmark
This is something i've also wondered for a long time, but i guess popularity might be an issue or Nintendo might complain you're using their games to market other products, i don't know.
The newest benchmark has been programmed by the devs themselves, I don't think you can be sued for that.
As mentioned above a legal and widely accessible means to do this wasn't availible until literally months ago thanks to delroth. Add to that the fact that few people even know this emulator exists and even fewer actually care enough to want to use it. What dolphin does, how it works, and the potential bottlenecks and optimizations employed by it are also not well understood making its usefulness as a general benchmark very limited because you can't draw clear conclusions from it (we still don't know why haswell is so fast with it for example).
The way delroth suggested is the first one to comare different cpus at all, but it is still not possible to include it in any script. I think there is a good chance to get ourself into some of such test suits, but there it's required to run without _any_ human interaction.
The lack of easy repeatable benchmark would by my guess though recording your own isn't difficult, but it's more work than putting timedemo demo1 in the console.
(02-24-2014, 12:14 PM)NaturalViolence Wrote: [ -> ]What dolphin does, how it works, and the potential bottlenecks and optimizations employed by it are also not well understood making its usefulness as a general benchmark very limited because you can't draw clear conclusions from it (we still don't know why haswell is so fast with it for example).
I don't know WTH Cinebench does or why I should care about it yet that's many people's weapon of choice. SuperPi is next to useless, but a few cling to it.
Well it just proves Haswell is much faster in single threaded performance and IPC over Ivy Bridge. Maybe no other app has tapped into Haswell's true potential because they are mostly multi-threaded based? Why do you think the hexa-core i7 extremes always come out on top? Surely, it's not about their single threaded performance.
Maybe haswell just improved on executing our shitty jit code.
Pages: 1 2 3