Main Menu

Assertion falied!

Started by Boga, September 27, 2010, 10:36:32 AM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

Boga

This is the error that occurs to me. Does anyone know why? Please Help!

Best regards,
Boga

Liviu Lalescu

#1
Please send me your input file and also give me the random seed at the start of generation for fet-5.14.2 for which you get crash (I would prefer a random seed which crashes in a shorter time - tell me that time and how many activities you reach until the crash).

You get crash each time you generate?

You should have reported this bug as soon as you met it.

Boga

The latest data are attached. I deleted the old data. :-(

Of 6-7 attempts, successfully be just one generation.

I immediately reported the error because I thought the problem in my computer.

Liviu Lalescu

#3
I tried your exact random seed and file on two different computers: my GNU/Linux, it was at max 1062 activities placed in around 1 hour and reached max 1063 5 minutes later. Now, it is at 1148 (max 1169), after 3 h 38 m.

On a Windows laptop, it just reached max placed 1063 after about 1 hour.

I don't know what to say. I am happy that it seems not to be a FET bug, maybe it is your computer.

Nobody reported such a bug this year, and I think there are many FET users who tried on many data sets.

Please, other users, could you try to solve the file of Boga using the starting random seed from the archive? Use the menu Settings->Advanced->Seed of random number generator, just before the generation start.

Boga, please inform us as soon as possible of updates to this problem (if you find out more details, or something changes, etc.).

Horatiu

I had the same error some time ago, on my laptop, quite a few times

It was a random error, as far s I could tell, and occured always after editing the .fet file with a text editor; usually after pasting a lot of data in the .fet file.

I did not find any errors in thefile, but I had to manually do everything. So, for me, was related to copy-pasting into the .fet file

Liviu Lalescu

QuoteI had the same error some time ago, on my laptop, quite a few times

It was a random error, as far s I could tell, and occured always after editing the .fet file with a text editor; usually after pasting a lot of data in the .fet file.

I did not find any errors in thefile, but I had to manually do everything. So, for me, was related to copy-pasting into the .fet file

Yes, you may get assert failed, when loading the .fet file or modifying it from FET or generating it, if the .fet file contains typographical mistakes. This is normal, as I did not have the patience to test all kind of possible mistakes in the .fet file.

This topic concerns generation bugs: assertions failed, for perfectly legal .fet files (especially those created from the FET menu). These problems are more important. Did you experience generation assert failed bugs?

A generation assert failed bug is a bug that appears in the process of generation (not while precomputing, which takes a very short period after you press "Start generation". Precomputing assert failed may be because of incorrect .fet data file).

Liviu Lalescu

#6
Boga, for your random seed, it solved on my computer (GNU/Linux) in approx. 10 hours. On another computer (laptop, Windows XP) it solved in approx. 12 hours (again with the same random seed).

Edited to add: I also tried with another random seed on my computer, it solved in about 9 hours (random seed at start: X=123, Y=123).

Boga

Several times I have successfully generated timetable.
I think the problem was in my computer, not in the FET.
FET is a great program.

Best regards
Boga

Liviu Lalescu

Are you sure it is exactly the same data that crashed FET previously? (I mean, for the exact file attached to this topic and random seed, FET crashed a time ago?)

Did you do something to fix your computer? (RAM change or other thing?)

Boga

Yes, the same files and random seed. Everything was OK.

Windows has made problems. I uninstalled some programs and cleaned the registry.

Brahim Montasir

I had encountered the same problem several times.I cleaned my computer 3times but with the same data it occurs again.But when I changed some constraints in my data, it runs normally.

Volker Dirr

If you get an assertion fail, please tell us (Liviu or me) the random seed number (you can find it in the results directory) and sent us your datafile. It is very difficult to help without that information.

Liviu Lalescu

#12
Yes, like Volker said, please give us:

1) FET version, operating system. I think you are referring to latest FET-5.14.5? Because in the past we solved some crash bugs in older FET versions (I think latest fixed crash bugs were about 1 year ago). If you have an older FET, please retry to see that the bug is fixed now or still existing.

2) The random seed at the start of generation (see it in FET -> Settings -> Advanced before starting the generation or in the results directory) and the input .fet file. If we have the same of these two, we will simulate exactly the same conditions/succession of events in the computer.

3) Did you generate/modify your .fet file by hand or with other program than FET? This is usually a cause of FET crashes. Additionally: did FET crash immediately when starting to generate or after more time, after placing some activities?

EDITED TO ADD number 4:

4) What type of constraints or combination of constraints do you think caused the crash?

Crash bugs in generation are critical and I am willing to fix them right away. Crashes when reading a wrong .fet file are not critical and I am not sure I will fix them, because there is too much work and they are not critical, in my opinion.

Brahim Montasir

Hi,
-I don't know exactly what constraint caused the assertion failed b/c I deleted all the constraints and put them anew.
-I used the 14.5 morocco version
-The assertion failed occured after the generation was complete, but couldn't get my timetable .
Best Regards

Liviu Lalescu

#14
Thank you for letting me know. So, it is a Morocco bug, and I know where it could be. But please try to remember:

1) Did you use other constraint types than in the example from 5_14_5-morocco?

2) What constraints types did you use (all of them) in your crashing file?

3) Maybe you can remember the file and line of the assert failed (this would be great).

4) The old Morocco 5_10_2 had the same bug? Or you didn't try with 5_10_2?