[comp.sys.hp] fortran problems on the 800's

Daniel.Stodolsky@cs.cmu.edu (01/17/90)

Folks - 

	this little fortran program dumps core on our 835's. Can anyone tell me
why?
If the array is dimensioned 1000000 instead, there is no problem  Thanks
in advance.

The code (compiled with no complier flags)

------------------------------------------ cut here
------------------------------------------
 real*8 rwork(1011117)
       rwork(1)=0.0d0
       stop
       end
------------------------------------------ cut here
------------------------------------------


		Dan Stodolsky
		Engineering Design Reserach Center

danner@cs.cmu.edu
"Happiness is an 835 Tubro SRX"

cunniff@hpfcso.HP.COM (Ross Cunniff) (01/19/90)

>	this little fortran program dumps core on our 835's. Can anyone tell
>me why? If the array is dimensioned 1000000 instead, there is no problem 
>Thanks in advance.

>The code (compiled with no complier flags)
...
>real*8 rwork(1011117)

I suspect you have your stack size set to 8Mbytes (8388608 bytes).
This implies that the largest real*8 array you could have would be
1048576 elements long (actually, it is probably a little smaller
since compilers often use the stack for temporaries).  You might
try compiling with -K, if you truly *need* an array that large,
or you might increase the stack size in the kernel (consult your
Systems Adminstrator's guide).

>		Dan Stodolsky
>		Engineering Design Reserach Center


				Ross Cunniff
				Hewlett-Packard Colorado Languages Lab
				...{ucbvax,hplabs}!hpfcla!cunniff
				cunniff%hpfcrt@hplabs.HP.COM