Error message killed in yje compilation of C Program

[gopalpathak]

Size of array in case of multiple declaration of array of higher size in C

--------------------------------------------------------------------------------

As per your suggestion the following program work

#include <stdio.h>
float x[9500][9500];

int main(void)
{

printf("fred\n");

x[9499][9499]=3.0/5.0;

printf("%f\n",x[9499][9499]);

return 0;

}

But when we declare more than one two-dimensional array then the error" KILLED" is flashed in the following prgram. Please try to help so that i could declare more than one two dimensional array of float type as it is required in our research program.

#include <stdio.h>
float x[9500][9500];
float y[9500][9500];
int main(void)
{

printf("fred\n");

x[9499][9499]=3.0/5.0;

printf("%f\n",x[9499][9499]);

y[9499][9499]=3.0/5.0;

printf("%f\n",y[9499][9499]);

return 0;

}

output message-"killed"

[wje_lf]

You previously raised this question here. The only new information you give us now is "yje". Maybe that's the clue. I googled yje, but didn't find anything useful.

What can you tell us about yje?

[nalg0rath]

Seems to work for me. Output:

fred
0.600000
0.600000

The problem might not be with the program but with the system you are trying to run it on. I'm not sure what the problem might be.

My guess is that when creating that big array the program allocates memory from the stack. There is probably a limit on your system somewhere that limits the amount of memory that can be allocated.

Edit:
You could test:

$ ulimit -a
$ ulimit -s 1024

The first command will display the current limit and the second is used to set the limit. I don't know how much memory you'll need. I'm currently at my school so I can't test it. On the computer I successfully ran the program on the limits are:

time(seconds)        unlimited
file(blocks)         unlimited
data(kbytes)         unlimited
stack(kbytes)        8192
coredump(blocks)     0
nofiles(descriptors) 256
vmemory(kbytes)      unlimited

> wje_lf: haha! Have you ever noticed that the "t" and "h" keys are very close to "y" and "j" ? It took me a while to figure that one out.

[drl]

Hi.

I agree with nalg0rath. If we are correct, then using ulimit, and perhaps a few experiments might reveal the maximum sizes of data arrays you can use. A parameterized version of your source file:

#include <stdio.h>

#define SIZE 40000
#define SM (SIZE-1)
float x[ SIZE ][ SIZE ];
float y[ SIZE ][ SIZE ];

int main( void )
{

  int mysize = SM;
  printf( "fred running with size %d \n", mysize );

  x[ mysize ][ mysize ] = 3.0 / 5.0;

  printf( "%f\n", x[ mysize ][ mysize ] );

  y[ mysize ][ mysize ] = 3.0 / 5.0;

  printf( "%f\n", y[ mysize ][ mysize ] );

  return 0;

}

On my system, one error message that I get from gcc for the SIZE 40000 on a box that has 1 GB of memory, gcc (GCC) 3.3.5 (Debian 1:3.3.5- is:

c1.c:5: error: size of array `x' is too large

For size 20000, I was able to compile without error, but for execution I got:

% ./c1
Killed

Note that this was during execution, not compilation as your thread title is written.

You can change the definition of SIZE to see what might fit in the computer in question.

For many questions like this, it is helpful to include something about your environment -- which OS and compiler you are using, the memory and CPU on your computer, etc. ... cheers, drl