[racket] DrRacket crashes

From: Joe Gilray (jgilray at gmail.com)
Date: Tue Feb 21 01:32:24 EST 2012

Hi Neal,

The code is not using tail recursion, but my test code is allocating and
deallocating large structures about 20 times before the crash.

I am limiting memory for DrRacket to 1GB (the system has 3GB).  DrRacket
was using about 800MB the last time I checked before the crash.  Note that
it doesn't give me the out of memory message.  When I hit Run, DrRacket
doesn't give back memory to the OS, but that is no surprise.

I know that this is not a very complete bug report.  I just put it out
there in case someone else is seeing something similar and maybe together
we can characterize the issue better.


On Mon, Feb 20, 2012 at 10:09 PM, Neil Van Dyke <neil at neilvandyke.org>wrote:

> Windows is very much running out of RAM, and is pounding swap?
> Were you able to look at the process sizes, so you know roughly how much
> virtual memory DrRacket was using, and whether any other processes were
> using lots of virtual memory?
> It could be that DrRacket was using a huge amount of virtual memory, or it
> could be that some other process was, but you were noticing it in DrRacket
> because a GC cycle was leading to thrashing.
> If DrRacket is indeed using a huge amount of memory, you might take a look
> at the code you were running in DrRacket, and whether all the memory usage
> of that code is necessary.  One of the first things to look at is whether
> you're using tail calls everywhere you can in recursion-intensive parts of
> the code.
> Also, I actually disable swap on Linux and Windows workstations I set up
> -- any process of mine that doesn't fit in RAM is dead to me.  My
> workstations are 5 year-old laptops with 2-3 GB of RAM each, and so far
> that has been more than enough RAM.
> --
> http://www.neilvandyke.org/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.racket-lang.org/users/archive/attachments/20120220/f688677a/attachment.html>

Posted on the users mailing list.