Re: [Maxima-discuss] gcl memory usage

[Gunter K. <gu...@pe...>]

Potential solution since with
https://gitlab.gnome.org/GNOME/gnome-shell/issues/64 a maybe similar
issue has hit the press:

Would it be reasonable to manually trigger the garbage collector
 - after executing a maxima command in cl-user::run () and
 - after running a testbench file
 - after each read_nested_list() and read_matrix() ?

This way we only trigger the garbage collector
 - if we loaded something potentially big (which doesn't happen too often)
 - and after a user-interaction. Garbage-collecting 10000 times in a
long maktelist tends to use much memory. But garbage-collecting once
after outputting maxima's output but before issuing the user prompt
should efficiently protect against fragmentation.

Rationale:
 - valgrind does tell we have no notable memory leaks
 - Without memory leaks and without wildly fragmenting the memory there
is simply no way that the testbench could ever exceed my system's memory.
 - The overhead from calling the garbage collector once in a while isn't
high
 - and triggering the gc between user interactions might be a smart way
of calling it only if something potentially big has happened.


@camm: Is there a way to manually trigger the garbage collector in gcl?

Kind regards,

  Gunter.