I would like to thank the help from the following friends:
Chuck Swiger <[EMAIL PROTECTED]>,Duncan Anker
<[EMAIL PROTECTED]>,Roman Neuhauser <[EMAIL PROTECTED]>,Greg
'groggy' Lehey <[EMAIL PROTECTED]>
for solving the memory limit problem.
Thanks very much!
Sincerely,
Frank
From: Chuck Swiger <[EMAIL PROTECTED]>
Hmm. Are you certain:
12-sec# dd if=/dev/zero of=/var/cache/bar bs=1024k count=64
64+0 records in
64+0 records out
67108864 bytes transferred in 18.790176 secs (3571487 bytes/sec)
13-sec# ls -l /var/cache/bar
-rw-r--r-- 1 root wheel 67108864 Jan 16 22:44
put your replies *below* the text you reply to. most questions@
subscribers are used to reading from top to bottom, from left to
right.
# [EMAIL PROTECTED] / 2003-01-15 22:02:49 -0600:
> >32-bit systems implementing VM typically could increase user-mode address
> >space up to 2 GB, al
Frank Li wrote:
[ ... ]
Cool! I created an additional swapfile according to its instruction.
Everything is smooth. Just one thing not quite understood.
# dd if=/dev/zero of=/usr/swap0 bs=1024k count=64
It actually created a 64GB swapfile.
Hmm. Are you certain:
12-sec# dd if=/dev/zero of=/
Hi, Duncan,
Thanks for your help!
From: Duncan Anker <[EMAIL PROTECTED]>
Yes, 2G - 1 is the maximum size a signed int can hold ... mind you, any
program needing more that 2G of memory really should be implementing
on-disk storage of data for itself.
Thanks for your suggestion. In this case I
On Fri, 2003-01-17 at 12:08, Frank Li wrote:
> Thanks for all of your replies,
>
> Now I can do it through recompiling the kernel and the limits did increase.
> I haven't tried whether it can increase over 2G (I would love that if it
> can!). The code I used cannot be easily changed to reduce
Thanks for all of your replies,
Now I can do it through recompiling the kernel and the limits did increase.
I haven't tried whether it can increase over 2G (I would love that if it
can!). The code I used cannot be easily changed to reduce memory
consumption but I think 2G would probably be en
On Fri, 2003-01-17 at 07:59, Frank Li wrote:
> Thanks, Roman,
>
> > > >options MAXDSIZ="(256*1024*1024)"
> > > >options MAXSSIZ="(256*1024*1024)"
> > > >options DFLDSIZ="(256*1024*1024)"
> > > >
> > > >Depending on what you're doing, you might well find using a 64-bit
> > >
Thanks, Roman,
> >options MAXDSIZ="(256*1024*1024)"
> >options MAXSSIZ="(256*1024*1024)"
> >options DFLDSIZ="(256*1024*1024)"
> >
> >Depending on what you're doing, you might well find using a 64-bit
> >platform (Alpha hardware? Solaris on SPARC?) to be more appropriate..
Frank Li wrote:
[ ... ]
Should I add more physical memory (if so should I reinstll OS)?
If you're actually doing something where the 512 MB datasize limit matters to
you, adding more physical memory will almost certainly speed things up.
No, don't reinstall; even Windows doesn't make you reins
On Wednesday, 15 January 2003 at 21:54:54 -0600, Frank Li wrote:
> Thanks, Greg,
>
>> From: Greg 'groggy' Lehey <[EMAIL PROTECTED]>
>>> Top shows:
>>> Mem: 180M Active, 21M Inact, 32M Wired, 13M Cache, 35M Buf, 656K Free
>>
>> This shows that you have much more than 128 MB of memory in the
>> machi
Thanks, Chuck,
Which file should I put in the following ?
Frank
32-bit systems implementing VM typically could increase user-mode address
space up to 2 GB, although variants on that and other things (ie, where
devices get mapped into memory) make that only an approximation. For
FreeBSD:
# C
Thanks, Greg,
From: Greg 'groggy' Lehey <[EMAIL PROTECTED]>
> Top shows:
> Mem: 180M Active, 21M Inact, 32M Wired, 13M Cache, 35M Buf, 656K Free
This shows that you have much more than 128 MB of memory in the
machine.
I checked it again (by rebooting it and seeing its booting messages, stupid
On Wednesday, 15 January 2003 at 18:36:31 -0600, Frank Li wrote:
> The situation is as follows:
>
> Physical memory is 128M, OS is FreeBSD 4.4.
> My C++ simulation code mallocs large amount of memory. When running, I
> found "top" shows SIZE is 514M, RES is 176M.
That's rather difficult on a mach
The situation is as follows:
Physical memory is 128M, OS is FreeBSD 4.4.
My C++ simulation code mallocs large amount of memory. When running, I found
"top" shows SIZE is 514M, RES is 176M. Then the code dumpped a core. The
code itself should have no problem. It's not complicated and it runs v
15 matches
Mail list logo