Well, there is something else. How is Axiom supposed to behave in the following case? (See axiom--main--1--patch-40/mnt/linux/doc/book.dvi Section 1.15.1)

--start axiom
top:
  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
17381 hemmecke  16   0 44072 8492 3016 S  0.0  1.7   0:00.12 AXIOMsys

a:=new(1000,1000,0.0)$Matrix(SF);
b:=new(1001,1001,0.0)$Matrix(SF);

Now top says
17381 hemmecke  17   0 44116  21m 3060 S  0.0  4.3   0:00.34 AXIOMsys

)undo -2

And top says
17381 hemmecke  16   0 44116  21m 3088 S  0.0  4.3   0:00.34 AXIOMsys

I understand that due to %% and )undo previous results must be cached, but why doesn't )undo release memory? Do I read the values of the top command incorrectly?

Ralf

Vanuxem Grégory wrote:
Hi,

Sorry,

Section 0.4.2 of axiom book => accessing earlier result.
It's a little dangerous in term of memory management...

Cheers,

Greg




-----Message d'origine-----
De : [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] la
part de Vanuxem Grégory
Envoyé : mercredi 22 juin 2005 03:10
À : Bill Page (E-mail)
Cc : axiom-developer@nongnu.org
Objet : RE: [Axiom-developer] Memory leak in Axiom


Hi,


-----Message d'origine-----
De : Page, Bill [mailto:[EMAIL PROTECTED]
Envoyé : mercredi 22 juin 2005 01:11
À : 'Vanuxem Grégory'
Cc : axiom-developer@nongnu.org
Objet : RE: [Axiom-developer] Memory leak in Axiom


On Tuesday, June 21, 2005 3:43 PM Vanuxem Grégory wrote:


It seems that there is a memory leak in Axiom;

Open a shell, type 'top' and observe residual and virtual memory.

Open AXIOMsys and type

new(1000,1000,0.0)$Matrix(SF);

and repeat this action.

I can not reproduce this error using the Windows version
of Axiom 0.1.4. This version of Axiom is based GCL 2.6.5.
On windows I can monitor memory usage with Task Manager,
Performance tab.

What version of Axiom and operating system are you using
for your test? What version of GCL was used to build Axiom?

GCL-2.6.6 on axiom patch 40


A (1000,1000) C double float array takes 7.63 Mb.

If I alloc a contiguous storage memory array of 1000x1000,
amount of virtual and residual memory augment of 10Mb. It seems
it is never released.

I haven't found this in gcl.

Could you describe the test you use on GCL? What version of
gcl are you using?

(make-array '(1000 1000) :element-type 'long-float :initial-element 0)

about 50 times


Any idea ?

If you repeat this process several times:


Sorry,
1) Open Axiom
4) display residual and virtural memory
-- avoid printing
new(1000,1000,0)$Matrix(SF);
-- memory has increased of 10Mb
new(1000,1000,0)$Matrix(SF);
-- memory has increased of 10Mb
new(1000,1000,0)$Matrix(SF);
-- memory has increased of 10Mb
new(1000,1000,0)$Matrix(SF);
-- memory has increased of 10Mb
new(1000,1000,0)$Matrix(SF);
-- memory has increased of 10Mb
new(1000,1000,0)$Matrix(SF);
-- memory has increased of 10Mb
new(1000,1000,0)$Matrix(SF);
-- memory has increased of 10Mb
.
.
.
a lot of times (axiom will crash) => xmalloc failed


Cheers,

Greg



For example do this 5 times in a row, do you find that the
total available memory is constantly decreasing?

If this is not that case, but only occurs on the first
iteration, then what you might be seeing is some kind of
executable cacheing being done by the operating system.
(Does linux do that? Some versions of Windows do. I see
such an effect under Windows XP.)

Regards,
Bill Page.





_______________________________________________
Axiom-developer mailing list
Axiom-developer@nongnu.org
http://lists.nongnu.org/mailman/listinfo/axiom-developer






_______________________________________________
Axiom-developer mailing list
Axiom-developer@nongnu.org
http://lists.nongnu.org/mailman/listinfo/axiom-developer



_______________________________________________
Axiom-developer mailing list
Axiom-developer@nongnu.org
http://lists.nongnu.org/mailman/listinfo/axiom-developer

Reply via email to