New VPX 10.8 Performance Issues?

[chonk_master]

A few months ago I updated my cab to the latest of everything for TWD. Since then I've noticed a memory leak type of performance issue and I've seen some similar posts.

 

If I power my cab on, fire up about any table and play, its fantastic as always. However after the table has been on for around 15-30 minutes or I've played a few tables, per the fps monitor F11, I get random massive lag spikes in response time (stutters), off the charts then back to normal. Once it starts it seems to persist over tables, through vpx restarts, etc. 

I've noticed similar posts especially in the latest VPW MM release, which is fing fantastic, but suffers this problem. The reason I think this is some kind of leak is because it persists and gets worse, but a reboot sets it back to normal.

 

I want to say its something in the vpin environment but if my memory isn't completely gone I don't think this issue started after the updates to the vpin software, I played TWD for hours after I got it running and don't recall this. So could it be a windows update? I'm on windows 11 and it is up to date? 

 

Some notes:

 

• This persists through various tables, I almost exclusively only have VPW/G5k/Flupper/Bigus1 tables, all of them suffer.
•  
• My 32bit client never worked after updating the vpin software to 10.8 stable release, I had to run a 32bit MS C++ update, I haven't done this for 64bit which I will now
•  
• Pup pack or no pup pack doesn't seem to make a difference, though I noticed once this stutter kicks in my topper flickers, but only when a table is active, desktop its fine no flickers on other monitors
•  
• I'm running pinvol in the background, haven't tried shutting this off but I will, though I'm doubting its the culprit as all this was working without issue
•  
• I hate to gut things because everything worked with no changes other than updating windows and the vpin environment

 

 

Any ideas before I deep dive into this?

[deadmanworking]

First things that come to mind is a software based issue or maybe thermal due to additional stress on CPU in the newer VPX, but hard to say without a few more details,  can you detail which versions of VPX, B2S, Vpinmame you are using exactly? 
Which edition of Windows? Also do you have latest Visual C runtimes installed?
Windows Defender exceptions set for all paths in use?
How are your thermals, can you run Openhardware monitor to see yours temps progressively and report back. 
And what is your cab setup anyway, which CPU, GPU, how many screens and resolution/refresh rate are we talking and how many intake / outflow fans and alignment of mobo / GPU inside

 

 

[chonk_master]

I did a little more play testing last night with VPW's MM and ST:TNG. The most telling piece of data I have at the moment is the F11 FPS display. Two indicators are the second row of real time numbers shows a average and max response time. The avg is around 1.2 with spikes in the 5-7ms range. This is typical but if I stare at it and just fling the ball around once the issue kicks in it spikes up to 70-90ms!!! The graph on the right reflects this and one line goes off the charts and then it goes back to normal. It also indicates its in the table script where the spike occurs. 

 

All this is making me think its another cpu contention issue. I have a few things to try today; one is install the c++ package for 64bit and I'm going to install pinaffinity and see if that makes a difference, I've had mixed results in the past.

 

deadman,

 

All my software is what is spec'd out in Flupper's TWD release. I did not install the december c++  package for 64bit. I did for the 32bit because vpx32 wouldn't run without the update.

 

Fairly certain its not a heat issue as the cab is currently in a cool location, but I'll keep this on the back burner. Cab is a 3070/i9 and gets a steady 240fps/1.2ms response all except when a spike occurs. It has 5 screens, 4k OLED playfield, CSD extreme kit. I love this thing and have to get it fixed lol.