[comp.unix.aix] 3rd party memory for RS-6000 530 installed!

guscus@katzo.rice.edu (gustavo e scuseria) (01/29/91)

We succesfully replaced our two 8 Mbytes memory boards with
4 Mbyte SIMMS provided by Impediment (617)-837-8877 (contact
Alex Sunguroff) in our ibm 530.
Now that we have 64 Mbytes, does anyone know how to configure
the kernel such that malloc will allocate 64 mbytes (or more) ?
Right now, malloc refuses to allocate more than 32 mbytes and
we have not been able to figure out how to change this.

--
Gustavo E. Scuseria
Assistant Professor
Department of Chemistry
Rice University

frank@leopard.Berkeley.EDU (01/31/91)

> Now that we have 64 Mbytes, does anyone know how to configure
> the kernel such that malloc will allocate 64 mbytes (or more) ?
> Right now, malloc refuses to allocate more than 32 mbytes and
> we have not been able to figure out how to change this.
>

I think that you need to look at the 'ulimit' command.  At least on my
machine, the limit for memory is 32M.

- Frank Feuerbacher


Disclaimer: I don't speak for my employer and they don't speak for me.

jfh@rpp386.cactus.org (John F Haugh II) (02/06/91)

In article <5055@awdprime.UUCP> ...@cs.utexas.edu:ibmaus!auschs!leopard.austin.ibm.com!frank writes:
>> Now that we have 64 Mbytes, does anyone know how to configure
>> the kernel such that malloc will allocate 64 mbytes (or more) ?
>> Right now, malloc refuses to allocate more than 32 mbytes and
>> we have not been able to figure out how to change this.
>
>I think that you need to look at the 'ulimit' command.  At least on my
>machine, the limit for memory is 32M.

You also need to be aware that the algorithm for AIX v3's malloc
is the same one used by BSD.  It uses "power of 2" allocation, which
causes requests to be rounded up to the next larger power of 2.  Thus,
for a 33MB request to be satisfied, you would have to have 64MB of
space remaining in your data segment.
-- 
John F. Haugh II                             UUCP: ...!cs.utexas.edu!rpp386!jfh
Ma Bell: (512) 832-8832                           Domain: jfh@rpp386.cactus.org
"I've never written a device driver, but I have written a device driver manual"
                -- Robert Hartman, IDE Corp.

richard@locus.com (Richard M. Mathews) (02/09/91)

jfh@rpp386.cactus.org (John F Haugh II) writes:

>You also need to be aware that the algorithm for AIX v3's malloc
>is the same one used by BSD.  It uses "power of 2" allocation, which
>causes requests to be rounded up to the next larger power of 2.  Thus,
>for a 33MB request to be satisfied, you would have to have 64MB of
>space remaining in your data segment.

On AIX V1.2 (PS/2 and 370) there is an alternate malloc supplied in
libmalloc.  This is a System V version of malloc which makes more
efficient use of address space at the expense of taking more time to
search for free space and possibly taking more page faults.  As with
AIX V3, the default malloc in libc is the one used by BSD.

As far as I know, V3 does not provide libmalloc.

Richard M. Mathews			D efend
richard@locus.com			 E stonian-Latvian-Lithuanian
lcc!richard@seas.ucla.edu		  I ndependence
...!{uunet|ucla-se|turnkey}!lcc!richard

jfh@rpp386.cactus.org (John F Haugh II) (02/09/91)

In article <richard.666055338@fafnir.la.locus.com> richard@locus.com (Richard M. Mathews) writes:
>On AIX V1.2 (PS/2 and 370) there is an alternate malloc supplied in
>libmalloc.  This is a System V version of malloc which makes more
>efficient use of address space at the expense of taking more time to
>search for free space and possibly taking more page faults.  As with
>AIX V3, the default malloc in libc is the one used by BSD.
>
>As far as I know, V3 does not provide libmalloc.

Nor could it, if IBM wanted to.  The malloc in the C library is 
pre-bound with the other modules in the shared library.  This would
give you two different "malloc()"'s, which won't work.  So far as
I know, you would have to not use any functions in the C library
in order to use your own malloc.

Or you could just rename the malloc you want to use to "mymalloc()"
and hope those two malloc's can co-operate with each other ...
-- 
John F. Haugh II                             UUCP: ...!cs.utexas.edu!rpp386!jfh
Ma Bell: (512) 832-8832                           Domain: jfh@rpp386.cactus.org
"I've never written a device driver, but I have written a device driver manual"
                -- Robert Hartman, IDE Corp.