| John IV |
|
IV/Play, MAME, MAMEUI
|
|
|
|
|
Reged: 09/22/03
|
|
Posts: 1970
|
|
Loc: Washington, USA
|
|
|
|
Send PM
|
|
|
Mame and Windows 11 22H2 slowdowns observations
#395002 - 10/22/22 01:34 AM
|
|
|
Just a PSA that a couple of new features of Windows 11 22H2 can really slow Mame down.
I saw 30%+ drops in some games of my bench suite (cmd line) when I recently ran it from the new Microsoft unified terminal which wraps up a new version of cmd.exe in it.
I found I had to run the cmd profile/instance as admin to return to the normal scores. This didn't use to be the case with regular windows console host version of cmd. Thankfully you can alter this default terminal application in the startup section and switch it back to windows console host instead of the new one.
The other slowdown is 'core isolation' is now defaulted to on in supported CPUs fresh installs or new PCs. You can turn this off in Windows Defender to regain the speed loss. This has impacted other games' fps as well.
https://www.pcworld.com/article/1069899/...by-default.html
Edited by John IV (10/23/22 04:37 AM)
|
john iv
http://www.mameui.info/
|
|
|
Re: Mame and Windows 11 22H2 slowdowns observations
[Re: John IV]
#395185 - 11/23/22 01:19 AM
|
|
|
Can you explain more how you got around this issue.
Can I just run Command Prompt as Administrator, or do I need to dig deeper.
Asking as I hope to have a chance to benchmark a i9-13900K system in the next couple of days, but it will be running Win 11.
|
If all else fails, Burn the manual.
|
|
| John IV |
|
IV/Play, MAME, MAMEUI
|
|
|
|
|
|
Reged: 09/22/03
|
|
Posts: 1970
|
|
Loc: Washington, USA
|
|
|
|
Send PM
|
|
|
Re: Mame and Windows 11 22H2 slowdowns observations
[Re: taz-nz]
#395186 - 11/23/22 02:06 AM
|
|
|
Hi Taz, two steps,
1. Launch the new terminal app.
2. Go to its settings (downward arrow on its titlebar/tab bar.
3. Choose Startup / Default Terminal Application and change it back to 'Windows Console Host'. This will return behaviour to prior CMD style and increase the performance.
---
Second thing is typing 'core isolation' in the windows search bar and go into it (settings / Defender) and disable 'memory integrity'.
That should return you to pre-22h2 speed levels. Note, I haven't re-tried w/ the above enabled again since I posted this thread, it's possible the performance pain has been mitigated since then in post 22h2 patches.
|
john iv
http://www.mameui.info/
|
|
|
MAME 0.249 Benchmark on i9-13900K
[Re: John IV]
#395196 - 11/25/22 08:08 AM
|
|
|
Thanks, system was on running Win 11 21H2, but I disabled core isolation and ran benchmark from Command Prompt.
Game i9-13900K
blitz 519.60%
crusnusa 564.25%
cubeqst 696.17%
cyvern 2246.35%
dkong 7659.62%
drivedge 682.31%
gauntleg 850.67%
gnbarich 2985.05%
gradius4 792.75%
gtfore06 537.56%
harddriv 1244.48%
kidniki 830.60%
kof98 2413.90%
mario 2008.34%
mk4 506.49%
pacman 21918.47%
pinkswts 3086.61%
pongf 836.03%
propcycl 463.66%
radikalb 450.58%
roadblst 2060.56%
robotron 11346.60%
rvschool 675.85%
scud 140.84%
sf2 4223.41%
sfa2 3396.95%
sfiii 2772.90%
slrasslt 878.43%
spacfury 322.98%
starblad 356.70%
starsldr 109.44%
sushibar 154.70%
tekken 1030.50%
tekken3 523.08%
vf2 302.02%
vfkids 268.36%
|
If all else fails, Burn the manual.
|
|
| John IV |
|
IV/Play, MAME, MAMEUI
|
|
|
|
|
|
Reged: 09/22/03
|
|
Posts: 1970
|
|
Loc: Washington, USA
|
|
|
|
Send PM
|
|
|
Re: MAME 0.249 Benchmark on i9-13900K
[Re: taz-nz]
#395199 - 11/25/22 11:49 PM
|
|
|
Nice, about a 21% uplift from my i7-12700K.
Were you on .249 release and what Ghz were you hitting on the P Cores, 5.4?
|
john iv
http://www.mameui.info/
|
|
|
Re: MAME 0.249 Benchmark on i9-13900K
[Re: John IV]
#395200 - 11/26/22 12:23 AM
|
|
|
HWinfo had the P cores averaging 5.475ghz and boosting to 5.785ghz for some parts of the run, I didn't have enough time with system to manually play core clocks.
|
If all else fails, Burn the manual.
|
|
|
Re: Mame and Windows 11 22H2 slowdowns observations
[Re: John IV]
#395233 - 11/30/22 05:32 PM
|
|
|
|
|
Previous 
Index 
Next 
Threaded 
Mame and Windows 11 22H2 slowdowns observations
Reply
[Re: 