Build Your Own Arcade Controls Forum

Software Support => GroovyMAME => Topic started by: adder on July 18, 2014, 06:13:38 pm

Title: bench tested a couple of games...
Post by: adder on July 18, 2014, 06:13:38 pm

Hi, i thought i would try a couple of benchmark tests with my groovymame (0153.015b) setup because i am using XP 64bit so wanted to check it was all installed/working ok, so i tried 3 different games (green beret, bad dudes, and sexy parodius), using both the 32bit and 64bit groovymame builds, on my XP 64bit PC (cpu is a core2duo E8400 3ghz)

i used 'mame <game> -bench 90' for all tests

Results:

sexyparo 32bit: Average speed: 171.20%
sexyparo 64bit: Average speed: 172.98%

gberet 32bit: Average speed: 5598.04%
gberet 64bit: Average speed: 5679.57%

baddudes 32bit: Average speed: 509.38%
baddudes 64bit: Average speed: 625.83%

Are these normal results? the reason i am paranoid is i remember reading about how one can expect a decent/noticeable increase in emulation speed between 32bit and 64bit mame builds. in reality, does that only apply to some of the drivers, or is there something amiss with my setup?

Title: Re: bench tested a couple of games...
Post by: Calamity on July 21, 2014, 07:22:01 am

Quote from: jadder on July 18, 2014, 06:13:38 pm

Are these normal results? the reason i am paranoid is i remember reading about how one can expect a decent/noticeable increase in emulation speed between 32bit and 64bit mame builds. in reality, does that only apply to some of the drivers, or is there something amiss with my setup?

Hi jadder,

I'd say the bigger difference is actually when using XP-32+MAME-32 vs XP-64+MAME-64, rather than MAME-32 vs MAME-64 on XP-64.

(in other words, it'd be a combination of the OS and MAME what performs better)