Yeah, if your working set is large than main memory, you're basically hosed.
Though it's arguably a flaw in the OS that it can't remain responsive enough
for the user to restore sanity even when there is a lot of swap and an
application is using a lot of it.
Richard: I think that, on windows, what you want to adjust is the size of the
page file. (It's just a different name for swap, but you may have better luck
searching for it under that name.)
On Sun, 4 Dec 2022, bill lam wrote:
When access locality does hold true, such as transposing a large matrix
in-place , the excessive disk swapping makes computers non responsive.
On Sun, 4 Dec 2022 at 7:56 AM Raul Miller <rauldmil...@gmail.com> wrote:
These days, a cheap laptop might have 8GB ram.
Why would 2-3GB ram be a better swap size here than a swap which is at
least as big as main memory?
(I guess we have write ups with advice about how much swap space is
useful, from back in the days when most machines had significantly
less than 1GB ram, and maybe we have similar write ups from people who
came up with some narrative describing why that advice is still
sometimes useful. But it seems to me that insufficient swap space
would lead to "unable to launch jbreak when J has grabbed too much
memory" issues.)
Thanks,
--
Raul
On Sat, Dec 3, 2022 at 5:58 PM bill lam <bbill....@gmail.com> wrote:
>
> Setting swap size equal physical ram is reasonable and is the default of
> some OS already. But here the problem is still too large to mitigate
> excessive disk swapping. Setting swap size to 2 or 3 GB is more
appropriate.
>
> That said, the issue is very serious on spinning hard disk but is
> acceptable when using ssd.
>
>
> On Sat, Dec 3, 2022, 11:10 PM Raul Miller <rauldmil...@gmail.com> wrote:
>
> > One approach here would be to have launched task manager before the
> > problem arises.
> >
> > If task manager is running and J is pre-selected within task manager,
> > selecting task manager (using the mouse or alt-tab) then hitting the
> > delete key (and then confirming on the resulting popup -- probably
> > hitting y or enter) should work.
> >
> > That said, I think you should seriously consider setting the swap file
> > size on this laptop to be as large as the amount of physical ram you
> > have. (Or, double that, if you have enough free space.)
> >
> > I hope this helps,
> >
> > --
> > Raul
> >
> > On Sat, Dec 3, 2022 at 9:10 AM Richard Donovan <rsdono...@hotmail.com>
> > wrote:
> > >
> > > Yes that is exactly what happens to me. I need a global switch to
throw
> > an interrupt if more than x cpu seconds elapses without termination.
Or,
> > when local storage is more than y% used. Preferably this could be
toggled
> > and set from an ide menu or from profile so I wouldn’t have to (forget
to!)
> > set it every time. There is nothing worse than having to reboot and
loses
> > loads of work!
> > >
> > > > On 3 Dec 2022, at 12:23, bill lam <bbill....@gmail.com> wrote:
> > > >
> > > > Sometimes this happens to me too when the process allocated far
more
> > memory
> > > > than physical ram available. The disk swapping for virtual memory
> > prevented
> > > > any mouse click or keyboard response. The problem only way to stop
it
> > is to
> > > > long pressing the power button.
> > > >
> > > > Therefore sometimes I want to set a lower limit of virtual memory.
It
> > is
> > > > better to trigger out of memory crash than waiting indefinitely and
> > finally
> > > > power off the computer.
> > > >
> > > >> On Sat, 3 Dec 2022 at 8:14 PM Richard Donovan <
rsdono...@hotmail.com>
> > wrote:
> > > >>
> > > >> Elijah;
> > > >>
> > > >> Most times I can’t even get in to task manager. As this problem
> > crashes
> > > >> the whole system and causes the need for a reboot, I am obviously
> > loathe to
> > > >> try to recreate it!ick Next time it happens I wrepost with more
info!
> > > >>
> > > >> Thanks
> > > >>
> > > >> Sent from Mail<https://go.microsoft.com/fwlink/?LinkId=550986>
for
> > Windows
> > > >>
> > > >> ________________________________
> > > >> From: Programming <programming-boun...@forums.jsoftware.com> on
> > behalf of
> > > >> Elijah Stone <elro...@elronnd.net>
> > > >> Sent: Friday, December 2, 2022 11:19:58 AM
> > > >> To: J Programming <programm...@jsoftware.com>
> > > >> Subject: Re: [Jprogramming] Avoiding J crash
> > > >>
> > > >> Rebooting is a bit drastic; can you not kill j from task manager?
> > > >>
> > > >> Beyond that, break _should_ work. I can look into making it a bit
> > more
> > > >> responsive for long array operations, but am curious to know what
> > > >> specifically
> > > >> has caused you trouble.
> > > >>
> > > >>> On Fri, 2 Dec 2022, Richard Donovan wrote:
> > > >>>
> > > >>> Hi!
> > > >>>
> > > >>> Is there any way to avoid the situation whereby by error I put J
> > into
> > > >> an infinite loop from which the only recovery is to reboot my
Windows
> > > >> laptop?
> > > >>>
> > > >>> I guess that my erroneous routine is locking the CPU to such a
such
> > an
> > > >> extent that I can’t even get to run Jbreak.
> > > >>>
> > > >>> I quite often receive a recoverable situation when a limit has
been
> > > >> exceeded but if I get the infinite loop situation I have lost my
> > entire
> > > >> session and have to start again.
> > > >>>
> > > >>> I really need a global setting which limits the amount of cpu
usage
> > and
> > > >> aborts a function before it gets out of hand.
> > > >>>
> > > >>> Thanks,
> > > >>>
> > > >>> Richard
> > > >>>
> > ----------------------------------------------------------------------
> > > >>> For information about J forums see
> > http://www.jsoftware.com/forums.htm
> > > >>
----------------------------------------------------------------------
> > > >> For information about J forums see
> > http://www.jsoftware.com/forums.htm
> > > >>
----------------------------------------------------------------------
> > > >> For information about J forums see
> > http://www.jsoftware.com/forums.htm
> > > >>
> > > >
----------------------------------------------------------------------
> > > > For information about J forums see
http://www.jsoftware.com/forums.htm
> > >
----------------------------------------------------------------------
> > > For information about J forums see
http://www.jsoftware.com/forums.htm
> > ----------------------------------------------------------------------
> > For information about J forums see http://www.jsoftware.com/forums.htm
> >
> ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
