Problem with test dynamics example

[tuyet]

Dear professor Taylor,
Dear Feap Admin,

I tested a dynamics example in Feappv.  This example ran well in Feappv version 2005.  However it cannot run in the new version of Feappv posted in your website.

There are two problems:

1. After the first inte command, I tried to plot mesh.  However Feappv quit the program immediately.

2. To solve the problem program the following command is used
batch         ! MACRO n.3: solve problem
dt,,.4
tran,newm      ! Transient analysis
loop,,200
loop,,4
time
loop,,10      ! Convergence loop
tang,,1
next
next
plot,wipe
plot,defo,1,1   ! Plot on deformed conf.
plot,fill      ! Plot filled mesh
plot,unde,,1   ! plot on undefor. conf.
plot,outl      ! Plot outlined mesh
next
end

But it has an error in output file:

**ERROR** Insufficient storage to allocate TANGS
           Required size =       17910
           Check data or choose other option.

This example also runs well in Feap version 8.2.

Could you please show me what is the problem of current version Feappv?

Thank you very much.  I am looking forward to hearing from you.

Best regards,

Tuyet

[Prof. R.L. Taylor]

Please post your input file so we can check what is happening.

[Prof. R.L. Taylor]

1. I have tested  your batch program with my input file and it works fine.

2. Have you solved other problems (static ones ) with no problems?  If so, do a static solution
on your mesh to see if the allocation is O.K.

3. Did you use solution command "CHECk" to make sure mesh is o.k.?

[tuyet]

Thank you very much for your response. 
I attach here the input files: Imens, Imens.mac, Imens.msh.  This example was introduced by L. Petrini & A. Reali as benchmark example for feappv.

Here are some problems

1. Check mesh command: it is fine.

2 Plot mesh after TIE command: it is fine
   Plot mesh after
       batch ! MACRO n.1: check mesh
       noprint ! Reduce output printing
       plot,defa,off
       plot,pers
       plot,hide
       plot,fill
       end
       0
       2000,300,2000
       0,0,1

      end

+ works in Feap v8.2
+ does not work under running my project in microsoft studio 8 + Intel
+ works under the built Feappv.exe

3. Static case:
+ works in Feap v8.2
+ does not work under running ''Start debugging'' icon under my project in microsoft studio 8 + Intel
+ works under the built Feappv.exe

4. Dynamics case:
+ works in Feap v8.2
+ does not work under running ''Start debugging'' icon under my project in microsoft studio 8 + Intel as well as the built Feappv.exe

Thank you very much.

My best regards,

Tuyet

[Prof. R.L. Taylor]

There are two known errors in memory allocation that probably affect your solution.  One is for proportional loads (which is probably the one affecting you now) and the other for compressed profile which will affect eigen-computations, etc.

A new version for the program should be posted soon and hopefully will fix your problem.

[tuyet]

Thank you very much for your quick reply.

I am looking forward to the new version.

[FEAP_Admin]

Tuyet, a new version has been uploaded to the FEAPpv web page.

[tuyet]

Thank you very much for your updated code. 

You program runs all right with the test problem posted in Feappv website.  Unfortunately the program cannot run well for dynamics in my computer.  I ran program in

1. Windows Xp, Microsoft Visual Studio 8 + Intel Fortrans in the first computer, which has the sam problem as I solve you before:

   Computing solution at time  4.0000E-01: Total proportional load  4.0000E-01
   Individual factors:
      1 =  4.0000E-01   2 =  0.0000E+00   3 =  0.0000E+00
 **ERROR** Insufficient storage to allocate TANGS
           Required size =       17910
           Check data or choose other option.

2. Windows 7, Microsoft Visual Studio 8 + Intel Fortrans Second computer, which happens another problem:

  Computing solution at time  4.0000E-01: Total proportional load  4.0000E-01
   Individual factors:
      1 =  4.0000E-01   2 =  0.0000E+00   3 =  0.0000E+00

 *ERROR* DCHECK: Attempt to input  1 value(s).  Input is:
              1

  Inconsistency occurred from INPUTS in data file Imens.mac                                                                                                                       
  at or near record number    65.  Input record is:

  end                                                                           

  If this record is correct error may result from
  missing blank record before new command type.

I don't know what happens right now. 

[Prof. R.L. Taylor]

Which input file are you using now? Send again please.

[tuyet]

Here are the attached files. Thank you very much.

[Prof. R.L. Taylor]

I am not able to test your program with a windows machine at this time.

The only thing I can see in your files relates to use of END records.  You have extra ones in the Imens.mac file (see attached).

I do not see how you can be getting the allocation problem unless you have a bad compile, that is you still have the old program being loaded.  Otherwise there is no way that I can understand receiving this message still.  I suggest you clean everything out, download the program again and recompile from the very beginning. 

[tuyet]

Thank you very much.

WIndows: I tried to recompile the new feappv program in windows and use new Imens.mac.  It happens the same problem.

Linux: With the new feappv program,  the program immediately interrupted after input the mesh.  In the feappv screen, it appears

feappv: malloc.c:2451: sYSMALLOc: Assertion `(old_top == (((mbinptr) (((char *) &((av)->bins[((1) - 1) * 2])) - __builtin_offsetof (struct malloc_chunk, fd)))) && old_size == 0) || ((unsigned long) (old_size) >= (unsigned long)((((__builtin_offsetof (struct malloc_chunk, fd_nextsize))+((2 * (sizeof(size_t))) - 1)) & ~((2 * (sizeof(size_t))) - 1))) && ((old_top)->size & 0x1) && ((unsigned long)old_end & pagemask) == 0)' failed.
Aborted (core dumped)

[FEAP_Admin]

tuyet,
I tried your problem on our Windows machine at it fails (though with a different error).
It seems that your problem is causing FEAPpv to generate a memory.  We will look into
it further and get back to you.

[tuyet]

Thank you very much, Feap Admin.

I am looking forward to hearing from you.

[FEAP_Admin]

A new version has been posted.  Your problem ran on our windows machine with the changes -- please give it a try now.