Jump to content

SAM Build with modular DMD drivers for pindmd1,2,3 and PIN2DMD


lucky1

Recommended Posts

23 minutes ago, peskopat said:

I used vpu sambuild 2.9 , I had this problem with SMB for a while , but when i launch with vp10.1 and an old vpinmame , no problem so i make a duplicate folder of Vp to play correctly with it .

But could you please try the official VPM 2.8, too, please? We have to somehow find out what is causing these stutters after all. So if its the compiler version used, or whatever.

(as i cannot reproduce it with any build at the moment on my setups)

Link to comment
Share on other sites

  • Replies 578
  • Created
  • Last Reply

I'll tried too  the official vpm2.8 before but stern table don't launch , and smb as the same problem , the only one to work is a vpm before september , i understand it's difficult to solve a problem you cannot reproduce , but a few people has it , and with many setup of course , very complex to say it's hadware but possibly .

Link to comment
Share on other sites

  • Content Provider

DJRobX has passed along his compiler settings which has some options set up differently from my settings. I'll run a build that should execute more like the build from 1216. This will hopefully help with WPC sound. It may not resolve issues with SMB if that has been going on since Sept. I should have that a little later today. Keep a look out in this thread.


Sent from my iPad using Tapatalk

Link to comment
Share on other sites

I asked him about his ddraw settings a while ago too.   :)

SMB looping problem is I believe a multi-faceted problem.   SMB is extra sensitive to CPU usage,    It will loop and repeat while heavier systems work fine and can "recover" with frameskipping.   There is probably a bug in the sound code somewhere that's contributing to this.  When folks reported problems with Kingpin I found all kinds of horribly broken buffer code and it became very stable once fixed.

GTS seems to commit audio to the sound library ahead of the "update" request.   That behavior is probably somehow involved, I haven't dug into MAME"s sound code further to understand how that's supposed to work 

I've also mentioned on the official VPM thread about the SwitchToThread code.  I believe this is related to why people with faster processors (and possibly VC2015 compiles) are getting distorted audio. 

Link to comment
Share on other sites

  • Content Provider

In theory, this should not make a difference but the DirectX kit that was documented with official is here:

DirectX7a SDK - ftp://ftp.physik.hu-berlin.de/pub/useful/dx7asdk/DXF/

and the DJRobX build links to Windows 8.1 SDK presumably for its integrated DirectX

I don't have the 8.1 SDK installed so I don't know if there is any difference here

I'm currently running a VC2010 build without the whole program opt and referencing DirectX7a

Link to comment
Share on other sites

i tested both ddraw settings with no issue , now i'm going to change all my setup , upgrade graphic card with a gtx970 and re-install w10 on ssd , So when it's ok i'll tell you if my problem is the same with my new hardware . thanks for both of you to search solving the smb sound sluttering , it's very king of you guys .

Link to comment
Share on other sites

  • Content Provider
3 hours ago, Blashyrk said:

I am trying to get the color rom for metallica to work, I made a folder named altcolor/mtl_164c/ in vpinmame dir.

and i renamed my mtl_164.dat to pin2dmd.pal and put it in the mtl_164c folder, but it's not working, Am I doing it right ?

Which dmddevice.dll do you use ? pin2dmd version is currently faster and has more features.

You also need to activate colorize option in pinmame (F1)

Link to comment
Share on other sites

 

16 minutes ago, lucky1 said:

Which dmddevice.dll do you use ? pin2dmd version is currently faster and has more features.

You also need to activate colorize option in pinmame (F1)

I use pin2dmd, ahh I didn't activate the colorize option, I'll try again tomorrow, thanks lucky1 

Link to comment
Share on other sites

31 minutes ago, Blashyrk said:

 

I use pin2dmd, ahh I didn't activate the colorize option, I'll try again tomorrow, thanks lucky1 

Also, the VPX-packaged version you mentioned above has coloring disabled. I'll release a new version shortly with coloring enabled again.

Link to comment
Share on other sites

  • Content Provider
2 minutes ago, freezy said:

Also, the VPX-packaged version you mentioned above has coloring disabled. I'll release a new version shortly with coloring enabled again.

Thank-you...I was going nuts wondering if I did something wrong....

Link to comment
Share on other sites

6 hours ago, DJRobX said:

I've also mentioned on the official VPM thread about the SwitchToThread code.  I believe this is related to why people with faster processors (and possibly VC2015 compiles) are getting distorted audio. 

I saw that, yes, but i wondered if simply removing it (so sleeping 1-2 ms 'too short') would lead to other issues in the end?!

Thats why i suggested there that you could maybe try a Sleep(0) or _mm_pause() instead and see how that goes on your machine?

All that stuff is so sensitive to the used hardware and OS version, so its hard to get right on all kinds of systems (which is really really stupid).

Link to comment
Share on other sites

6 hours ago, CarnyPriest said:

In theory, this should not make a difference but the DirectX kit that was documented with official is here:

DirectX7a SDK - ftp://ftp.physik.hu-berlin.de/pub/useful/dx7asdk/DXF/

and the DJRobX build links to Windows 8.1 SDK presumably for its integrated DirectX

I don't have the 8.1 SDK installed so I don't know if there is any difference here

I'm currently running a VC2010 build without the whole program opt and referencing DirectX7a

These newer SDKs do not feature DX7 anymore, thats why this old thing is still needed to build VPM, unless somebody finds the time to also port this to DX9.

Or did i miss something there?!?

Link to comment
Share on other sites

  • Content Provider

Miss something? I'm not sure. As Rob mentioned to me: 

"To get it working on VC2010 the biggest "trick" was pointing the VC directories at the Windows 8.1 SDK files.  It doesn't seem to be able to pick that up automatically like VC2015 does.   The built in Windows SDK is missing Direct3d 7." 

Just presuming that he's using some different set than the old DXF off of that FTP. Again, this is probably unrelated to the sound issue. 

 

 

Link to comment
Share on other sites

1 hour ago, toxie said:

I saw that, yes, but i wondered if simply removing it (so sleeping 1-2 ms 'too short') would lead to other issues in the end?!

If you put a 1ms sleep in a loop (not the "early" sleep time), it can never short sleep.   It can only oversleep.    If it oversleeps, the next delay loop will factor that in and it will balance out.     I've been testing this theory on my cab for weeks, not a single issue or sound blip.   I've concluded that VPM does not need that delay loop to be super precise.   Given that, I don't think there's a need to waste cycles trying to hone in on sub-1ms precision delays.   

 

Link to comment
Share on other sites

14 hours ago, CarnyPriest said:

Ok, here's a VC2010 build. At least it loads twd_156h. Please let us know about WPC sound. Seems like users have had issues with Gilligan's Island, for example.

 

https://dl.dropboxusercontent.com/u/45430846/VPinMAME.dll

Hi Carny,

I've quickly tested this build this morning and could launch multiple SAM ROMs without any issue.

I've launched 4 or 5 SAM ROMs from VPM setup test and launched 3 times Spider-Man VE from VPX and didn't notice any freeze of my setup.

Thanks ;)

++

Link to comment
Share on other sites

8 hours ago, CarnyPriest said:

Miss something? I'm not sure. As Rob mentioned to me: 

"To get it working on VC2010 the biggest "trick" was pointing the VC directories at the Windows 8.1 SDK files.  It doesn't seem to be able to pick that up automatically like VC2015 does.   The built in Windows SDK is missing Direct3d 7." 

Just presuming that he's using some different set than the old DXF off of that FTP. Again, this is probably unrelated to the sound issue.

Weird. At least for me the 8.1 SDK does not feature the DX7 files.

Link to comment
Share on other sites

7 hours ago, DJRobX said:

If you put a 1ms sleep in a loop (not the "early" sleep time), it can never short sleep.   It can only oversleep.    If it oversleeps, the next delay loop will factor that in and it will balance out.     I've been testing this theory on my cab for weeks, not a single issue or sound blip.   I've concluded that VPM does not need that delay loop to be super precise.   Given that, I don't think there's a need to waste cycles trying to hone in on sub-1ms precision delays.

True, i forgot that you also disabled the >2ms check in your snippet. So if it does not lead to problems so far, why not give it some public tests then? I just did not want to include it into 2.8, as it was not tested yet by a lot of people, and all my reading up on the general topic of putting threads to sleep was a nightmare so far. But you understand all the VPM logic there much better than me by now.  :)

In case of VP it will also not matter, as its sole use is just limiting the framerate at the moment. So i'll change that, too.

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

×
  • Create New...