EmuNewz Network
benchmarking and rpcs3 - Printable Version

+- EmuNewz Network (https://www.emunewz.net/forum)
+-- Forum: Official Boards for Emunews Affiliated Emu Projects (https://www.emunewz.net/forum/forumdisplay.php?fid=47)
+--- Forum: Official RPCS3 Forum [archive] (https://www.emunewz.net/forum/forumdisplay.php?fid=172)
+---- Forum: RPCS3 - Support & Issue Reporting (https://www.emunewz.net/forum/forumdisplay.php?fid=163)
+---- Thread: benchmarking and rpcs3 (/showthread.php?tid=164426)



benchmarking and rpcs3 - ZeCoxao - 09-28-2014

would it be possible to have a thread for benchmarking results? something like, use same sample X and post results (build , OS, specs, etc)?


RE: benchmarking and rpcs3 - Ekaseo - 09-28-2014

(09-28-2014, 12:32 PM)ZeCoxao Wrote: would it be possible to have a thread for benchmarking results? something like, use same sample X and post results (build , OS, specs, etc)?

there is a test project where the devs test their features, https://github.com/DHrpcs3/ps3autotests , and the ppl generaly post what speeds they get on a game (+ screenshot) and most of them have their computer specs written in their signatures.


RE: benchmarking and rpcs3 - ZeCoxao - 09-28-2014

ok. well, what i mean was more of a specific thread. most of the information regarding benchmark is spreaded all over the forums. i was aiming at something more specific.
edit: or we can keep it as it is, i dunno


RE: benchmarking and rpcs3 - flashmozzg - 09-28-2014

I don't think it's very good idea as performance is quite unstable and could vary greatly between different builds or even runs.


RE: benchmarking and rpcs3 - ZeCoxao - 09-28-2014

ok, then forget about it Smile


RE: benchmarking and rpcs3 - tambre - 09-28-2014

(09-28-2014, 02:51 PM)ZeCoxao Wrote: ok, then forget about it Smile

I mean it doesn't sound bad...
Like make test program that runs different calculations and functions.
Then make it run them like 10 times and calculate the average?


RE: benchmarking and rpcs3 - flashmozzg - 09-28-2014

(09-28-2014, 03:10 PM)tambre Wrote:
(09-28-2014, 02:51 PM)ZeCoxao Wrote: ok, then forget about it Smile

I mean it doesn't sound bad...
Like make test program that runs different calculations and functions.
Then make it run them like 10 times and calculate the average?

Well it would be pointless that way. People can as well just run CatZilla or 3DMark. And emulation performance varies greatly right now so results posted soon become obsolete. Right now compatibility is the main goal.