338 replies to this topic

[toxie]

@RipleYYY: Would be neat if you could track down what causes the stutters a bit more (i.e. certain table elements or stuff like that)..

[RipleYYY]

well, i dont know... thats the prob

 

the "feeling" of ball movement is perfect on 9.1.4

its not on 9.1.5 445 ... on various tables as indy500, afm, scared stiff

 

as said, will try rel 474 today, and even install last nvidia drivers 314.07

 

VP is wonderfull for sure, but works different for evbd of us (lol inside of course) and always needs personnal tweaks

 

:-)

[toxie]

Okay, if it's hard to track down that way, maybe you could also try older versions of koadics builds and tell which build version was the first to show the hickups, so that we could at least try to get to a rough timeframe where things got worse for you..

[RipleYYY]

at the moment only tried 445

 

well, as it seems i'm the 1st and only one to notice that, dont worry too much (even if i'm happy if you take care of course)

as i see no difference with tables less heavy (lost world, harlem gtt and similar), i will say its my "weak" system in cause... of course linked with the new version !?

 

the question should than be : is it possible that this new version requires more "power" than the 9.1.4 ? or i'm just (a little) mad

 

ok for now, more later...

 

++

[toxie]

Actually it should be the other way round (9.1.5 slightly faster), as we included some more optimizations, so that's why it would be good to track this down, maybe we included some bug somewhere..

So please just test some versions of koadic so that we could pinpoint this better, which version introduced this..

[tttttwii]

I experienced exactly the same. I played Atlantis FS and as the ball droped in, it left some marks on the table, which were not removed after launching the ball (please see pic). This effect disappears, if I switch back to older vpinball.exe. It happened not in some other tables (eg AFM).

Attached Files

•  Atlantis.JPG   167.89KB   73 downloads

[RipleYYY]

and than :

 

9.1.5 rev 474, same feeling as 445

9.1.5 rev 474 with new nvidia 314.07, same as 445... + color bug (fade) while playing VP, but OK on the desktop (thats the driver for sure)

9.1.5 rev 474 with old nvidia 270.61, same as 445

and 9.1.4 with 270.61, again feel better, but may be not as was thinking before !?

 

conclusion after spending houres on the pincab : i'm tired, and quite surelly... mad, as i see balls of steel everywhere !!!

 

now, this is visual, i'm not using FPS software

 

...

Edited by RipleYYY, 12 March 2013 - 04:35 PM.

[The Loafer]

Rip: Toxie did not want you to test video card drivers, but other vp.exe version's found in koadic's signature. You should be able to get all the testing done in 15 minutes nice you have downloaded the other .exe's

[open6l]

Question - any thought of having Quit to Editor as the default option when existing a table - would help cab owners from grabbing a keyboard everytime they want to exit a table manually (whether editing or testing a table, etc...)...

 

 

should add that I don't have Q mapped as exit - I have ESC (unfortunately I'd have to open up my cabinet and re-wire my Exit button from ESC to Q since my iPac 2 doesn't support remapping of the ESC key).

 

If not then I guess I'll have to rewire at some point.

Edited by open6l, 13 March 2013 - 12:22 AM.

[jpsalas]

Okay, if it's hard to track down that way, maybe you could also try older versions of koadics builds and tell which build version was the first to show the hickups, so that we could at least try to get to a rough timeframe where things got worse for you..

 

I have an i7 running at 3,6ghz and an Nvidia GTX 680, and the last version in which the ball moves perfectly is the 442, after that the ball is redrawn badly while playing. I didn't notice this before because I used the newer revisions only to build the table I'm working on, and everything seemed to be ok. But using the Star Wars trilogy as a test shows a very bad ball redraw after rev 442, it is like the ball is not completely redrawn, and I can see the ball disappearing and appearing as in steps. I could not notice that in a table without alpha ramps or primitives.

[RipleYYY]

haaaaaaayaaaaaaaaaa JP !

 

i'm not more alone (lol), and than this means i'm not... mad ???

 

[toxie]

@JP: Thanks, that should help! One more thing though: What is the exact later version you tested that showed problems first? 443 or larger?

[jpsalas]

There is no 443, so the first one with a bad ball redraw is 444

[toxie]

Very weird.. So this is related to the primitive performance fixes and/or that unified code of the transformation matrices..

Both don't make sense to cause that issue..

 

I'll investigate..

[Pinball999]

If it can help, I also had the feeling that the latest builds introduced a bit of stutter, but it's now gone with 474..

EXCEPT when I'm using the Adaptive Vsync (on T3 for instance), in that case, the ball sometimes completely freezes for a short time.

[htamas]

Using adaptive vsync completely messes up VP for me.

Interestingly, the effect is the opposite for FP, I have to check vsync in FP video options to get smooth ball movement.

[jpsalas]

If it can help, I also had the feeling that the latest builds introduced a bit of stutter, but it's now gone with 474..

EXCEPT when I'm using the Adaptive Vsync (on T3 for instance), in that case, the ball sometimes completely freezes for a short time.

Try my table Star Wars trilogy. I made that table using several alpha ramps and primitives, just to see what VP914 could do. So it is a nice table to test the latest builds. I have no stutter at all using the VP914, but the ball stutters from version 444 onwards. It is not that it stutters, it is more like it is not drawn completely, and it drop frames, so the ball moves jumping from one point to another, even the flippers are sluggish with the primitives on top.

[Pinball999]

If it can help, I also had the feeling that the latest builds introduced a bit of stutter, but it's now gone with 474..

EXCEPT when I'm using the Adaptive Vsync (on T3 for instance), in that case, the ball sometimes completely freezes for a short time.

Try my table Star Wars trilogy. I made that table using several alpha ramps and primitives, just to see what VP914 could do. So it is a nice table to test the latest builds. I have no stutter at all using the VP914, but the ball stutters from version 444 onwards. It is not that it stutters, it is more like it is not drawn completely, and it drop frames, so the ball moves jumping from one point to another, even the flippers are sluggish with the primitives on top.

You meant VP915, right?

I've compared SWT with 442 and 474, and to be honest, if there is a difference I'm not able to spot it.

The only instance when the ball seem redrawn, is when I'm using adaptive vsync.

 

But I've made so many changes on my cab since a few weeks, that it has become really difficult to nail something precisely.

For instance, I'm using the ledcontrol.vbs from that thread:

http://www.vpforums....287#entry214045

Other possible factors that could create some differences:

4gb patch on all builds

560ti set to quality and not performance

Alpharamp cursor one mark before middle.

[toxie]

If it can help, I also had the feeling that the latest builds introduced a bit of stutter, but it's now gone with 474..

EXCEPT when I'm using the Adaptive Vsync (on T3 for instance), in that case, the ball sometimes completely freezes for a short time.

Try my table Star Wars trilogy. I made that table using several alpha ramps and primitives, just to see what VP914 could do. So it is a nice table to test the latest builds. I have no stutter at all using the VP914, but the ball stutters from version 444 onwards. It is not that it stutters, it is more like it is not drawn completely, and it drop frames, so the ball moves jumping from one point to another, even the flippers are sluggish with the primitives on top.

 

This is some really really weird stuff.. If i test the builds from koadic on my machine, i actually get this effect (on T2 Chrome f.e.) between 467 and 470!

But if i revert the changes done in this range on the latest revision, it's still there! WTF?!?!

I fear that this is a completely unrelated bug that just got triggered by some changes done somewhere else..

[blur]

can you make a diff to see all changes from 442 to 444?