"Direct solver solution was non-finite; Trying to fix..."

If you have problems installing or running Algodoo, have found a bug or need in-game help - this is it!

"Direct solver solution was non-finite; Trying to fix..."

Postby mold999 » Mon Sep 30, 2013 11:48 pm

A long time ago i had this issue with scene with a lot of hinges, where the objects the hinges were attached to would not move or react to physics at all. If I deleted one of the object with a hinge, then the entire thing would spaz-out uncontrollably. I checked the console, and got this error:

"Direct solver solution was non-finite; Trying to fix..."

I'm not sure what it is that is not finite, I checked my scene, couldn't find anything that was infinite where it shouldn't have been.

Not sure how to change the solver, or even if there is a way.

Any ideas?
Phundementalism?

Ain't nobody got
thyme fo dat.
User avatar
mold999
 
Posts: 225
Joined: Sun Jan 02, 2011 1:35 am
Location: [176436, 3763234]

Re: "Direct solver solution was non-finite; Trying to fix..."

Postby electronicboy » Tue Oct 01, 2013 2:02 am

That's more an internal error with the physics calculation itself.

If I remember correctly, the Direct solver is for hinges, which means that in the hinge menu, you should be able to drop the legacy version a number to use the older system.
When asking for help, READ THE STICKIES!
electronicboy
 
Posts: 1694
Joined: Mon Aug 31, 2009 6:18 pm

Re: "Direct solver solution was non-finite; Trying to fix..."

Postby mold999 » Tue Oct 01, 2013 2:10 am

Yes, I also have no idea what the "legacyMode" thing does. Does it affect the hinge solver?
Phundementalism?

Ain't nobody got
thyme fo dat.
User avatar
mold999
 
Posts: 225
Joined: Sun Jan 02, 2011 1:35 am
Location: [176436, 3763234]

Re: "Direct solver solution was non-finite; Trying to fix..."

Postby electronicboy » Tue Oct 01, 2013 11:08 am

The "LegacyMode" setting is a variable which tells the physics system what solver to use.
That setting can tell Algodoo to use the older solver, which is why most scenes still work in the newer versions.

You should be able to drop that variable down one, and Algodoo should pass all calculations to the older solver, which may or may not help your issue.
When asking for help, READ THE STICKIES!
electronicboy
 
Posts: 1694
Joined: Mon Aug 31, 2009 6:18 pm

Re: "Direct solver solution was non-finite; Trying to fix..."

Postby mold999 » Tue Oct 01, 2013 11:54 pm

Yes, I did that, and now everything's working fine. Thanks!
Phundementalism?

Ain't nobody got
thyme fo dat.
User avatar
mold999
 
Posts: 225
Joined: Sun Jan 02, 2011 1:35 am
Location: [176436, 3763234]

Re: "Direct solver solution was non-finite; Trying to fix..."

Postby Xray » Thu Oct 03, 2013 6:46 am

Is there a "rule of thumb" for determining which direct solver to use, or is it always a case of trial-and-error?
User avatar
Xray
 
Posts: 500
Joined: Sun Jun 17, 2012 6:12 am
Location: USA

Re: "Direct solver solution was non-finite; Trying to fix..."

Postby electronicboy » Thu Oct 03, 2013 1:58 pm

Essentially, always use the direct solver, it was created to provide a more realistic axel, attempting to remove some of the issues of stretching hinges in the older versions of Algodoo. In the older system, algodoo always tried to pull the two hinged objects together, even if it meant breaking the process of the collion system in the progress.

The newer system is more respective of the collion system, and allows hinges to stretch a little if it's impossible to pull them together, which also boosts performance because of this, as well as also employing other techniquies to remove the energy loss caused by stretching hinges.

I'm sure that there are probably more reasons why the new solver was made, but essenially, you should always use the new solver, unless it's causing an issue for you. The newer solver is awesome, however, it doesn't cope as well as the older system with conditions it shouldn't be expecting. There are falls and advantages of the older and newer system, you just have to use the one which works for you.

PS, the newer system is the Direct solver, and, I think that older hinges where processed in the main algodoo physics system, as opposed to using a different set of formulas and code.
When asking for help, READ THE STICKIES!
electronicboy
 
Posts: 1694
Joined: Mon Aug 31, 2009 6:18 pm

Re: "Direct solver solution was non-finite; Trying to fix..."

Postby Xray » Thu Oct 03, 2013 7:01 pm

Okay, Thanks! :thumbup:
User avatar
Xray
 
Posts: 500
Joined: Sun Jun 17, 2012 6:12 am
Location: USA


Return to Help / Bug reports

Who is online

Users browsing this forum: No registered users and 10 guests

cron