array size limitation in ifc

array size limitation in ifc

hello,

ifc gives strange problems when the array size increases beyond some limit. what is the limit exactly? is there a way to ask it to accommodate larger array sizes?

thanks,
michelle.

7 posts / 0 new
Last post
For more complete information about compiler optimizations, see our Optimization Notice.

I would expect that any limit is determined by the operating system and perhaps object language, and not some sort of compiler-specific limit that can be "adjusted". What sort of limit and strange errors are you seeing?

Steve

Steve - Intel Developer Support

the operating system is linux version 2.4.9 (red hat 7.1.2). the processor is a xeon 2.2 ghz processor. the strange error that i get is that the system libraries are not linked if i increase the array size as shown by the following output to the 'ldd' on the executable:

tahoe:~ # ldd x_ion.exe
not a dynamic executable
tahoe:~ #

i have a 12 complex*16 arrays of size (max_arr,max_arr), one real*8 array of size (max_arr,max_arr), and an integer vector of size (max_arr).

it works fine if max_arr is less than or equal to 2206. i get the error when max_arr >= 2207.

this is the output to the ldd command when max_arr <= 2206:

tahoe:~ # ldd x_ion.exe
libm.so.6 => /lib/libm.so.6 (0x40017000)
libcxa.so.3 => /opt/intel/compiler70/ia32/lib/libcxa.so.3 (0x4003a000)
libc.so.6 => /lib/libc.so.6 (0x4006c000)
/lib/ld-linux.so.2 => /lib/ld-linux.so.2 (0x40000000)

please help me out.

thank you so much.
michelle.

> I would expect that any limit is determined by the
> operating system and perhaps object language, and not
> some sort of compiler-specific limit that can be
> "adjusted". What sort of limit and strange errors
> are you seeing?
>
> Steve

Looks like you're crossing over a 1GB boundary. I am not familiar enough with Linux to know if that's reasonable, but it's not the compiler. You don't say what error you get when you increase the size. Have you tried "limit datasize unlimited" or "ulimit" to see if that helps?

Steve

Steve - Intel Developer Support

steve,

> Looks like you're crossing over a 1GB boundary. I am
> not familiar enough with Linux to know if that's
> reasonable, but it's not the compiler. You don't say
> what error you get when you increase the size.

the program runs for a while and then it says:

Address error

End of diagnostics

> Have you tried "limit datasize unlimited" or "ulimit" to
> see if that helps?

the datasize is already unlimited as can be seen from the output of the limit command:

tahoe:~# limit
cputime unlimited
filesize unlimited
datasize unlimited
stacksize unlimited
coredumpsize 0 kbytes
memoryuse unlimited
vmemoryuse unlimited
descriptors 1024
memorylocked unlimited
maxproc 7168

- michelle

If you wish to experiment, and a fully static link is possible, that might gain you some more data space. The address space available below the point where .so libraries are loaded would be a function of linux, not of the particular compiler in use.

> If you wish to experiment, and a fully static link is
> possible, that might gain you some more data space.
> The address space available below the point where
> .so libraries are loaded would be a function of
> linux, not of the particular compiler in use.

would you tell me how to increase this space in redhat 7.2?

thanks!
michelle.

Leave a Comment

Please sign in to add a comment. Not a member? Join today