I just build 6.3.beta4 and when building the doc, I noticed large memory
usage spikes when compiling the "master document"s. I ended up eating up
all of my RAM (4 GB) plus swap space (2 GB) when running it in parallel (4
cores/threads) and it was trying to build 3-4 different master documents.
On 2014-06-18, kcrisman wrote:
>
>
> On Tuesday, June 17, 2014 10:41:39 PM UTC-4, Travis Scrimshaw wrote:
>>
>> Apparently this is related to Gap running out of memory and will (should)
>> be fixed with the merging of #16440. See also
>> https://groups.google.com/forum/#!topic/sage-devel/4r0Quzz
On Tuesday, June 17, 2014 10:41:39 PM UTC-4, Travis Scrimshaw wrote:
>
> Apparently this is related to Gap running out of memory and will (should)
> be fixed with the merging of #16440. See also
> https://groups.google.com/forum/#!topic/sage-devel/4r0QuzzDlg0.
>
> Best,
>
Possibly... but I hav
Apparently this is related to Gap running out of memory and will (should)
be fixed with the merging of #16440. See
also https://groups.google.com/forum/#!topic/sage-devel/4r0QuzzDlg0.
Best,
Travis
On Tuesday, June 17, 2014 7:11:30 PM UTC-7, kcrisman wrote:
>
> FWIW, I also run into this occasi
FWIW, I also run into this occasionally and have no idea how to fix it.
It's particular annoying when trying to review tickets.
http://trac.sagemath.org/ticket/14669 doesn't seem to have made it worse
or better, has it?
--
You received this message because you are subscribed to the Google Gr
Dima Pasechnik wrote:
On 2014-06-01, leif wrote:
leif wrote:
[...]
FWIW,
find src/doc/ -type d -empty -exec rmdir {} \;
make doc
solved that error, but I'm of course not sure whether removing empty
folders was necessary, or just the second attempt then worked.
It does not help my
On 2014-06-01, leif wrote:
> leif wrote:
[...]
>
> FWIW,
>
>find src/doc/ -type d -empty -exec rmdir {} \;
>make doc
>
> solved that error, but I'm of course not sure whether removing empty
> folders was necessary, or just the second attempt then worked.
It does not help my totally f-d i
leif wrote:
leif wrote:
John Cremona wrote:
On 1 June 2014 17:33, leif wrote:
FWIW, little "update":
This is what I got after
make doc-clean && ./sage --docbuild --no-pdf-links reference html
...
writing output... [ 93%] sage/combinat/tutorial
writing out
leif wrote:
John Cremona wrote:
On 1 June 2014 17:33, leif wrote:
FWIW, little "update":
This is what I got after
make doc-clean && ./sage --docbuild --no-pdf-links reference html
...
writing output... [ 93%] sage/combinat/tutorial
writing output... [ 93%]
John Cremona wrote:
On 1 June 2014 17:33, leif wrote:
leif wrote:
Dima Pasechnik wrote:
Is there anything short of "make distclean && make" to fix the stuck
docbuild:
By "stuck" I mean that I run make and
the output (and logs/dochtml.log) getting stuck at
[history_a] no targets are o
On 1 June 2014 17:33, leif wrote:
> leif wrote:
>>
>> Dima Pasechnik wrote:
>>>
>>> Is there anything short of "make distclean && make" to fix the stuck
>>> docbuild:
>>>
>>> By "stuck" I mean that I run make and
>>>
>>> the output (and logs/dochtml.log) getting stuck at
>>>
>>> [history_a] n
leif wrote:
Dima Pasechnik wrote:
Is there anything short of "make distclean && make" to fix the stuck
docbuild:
By "stuck" I mean that I run make and
the output (and logs/dochtml.log) getting stuck at
[history_a] no targets are out of date.
[structure] loading cross citations... looking
Volker Braun wrote:
On Thursday, May 29, 2014 9:45:52 PM UTC+1, leif wrote:
Well, nevertheless, AFAIK it uses Python's multiprocessing even with
SAGE_NUM_THREADS=1, just like 'sage -b' does.
Whats wrong with that?
As Dima pointed out, it doesn't only (needlessly) require specific
re
On Thursday, May 29, 2014 9:45:52 PM UTC+1, leif wrote:
>
> Well, nevertheless, AFAIK it uses Python's multiprocessing even with
> SAGE_NUM_THREADS=1, just like 'sage -b' does.
>
Whats wrong with that? You want an extra untested codepath to handle that
special value?
--
You received this mess
Volker Braun wrote:
The doc builder respects the usual SAGE_NUM_THREADS environment variable.
Well, nevertheless, AFAIK it uses Python's multiprocessing even with
SAGE_NUM_THREADS=1, just like 'sage -b' does.
I've been considering the latter a bug for years, but others apparently
didn't car
On 2014-05-29, Volker Braun wrote:
> If its really because you are running out of memory then the OOM killer
> will send a SIGTERM, do you see that in strace?
>
> The doc builder respects the usual SAGE_NUM_THREADS environment variable.
Really? With
$ SAGE_NUM_THREADS=1 make
I still get *two*
On 2014-05-29, Volker Braun wrote:
> If its really because you are running out of memory then the OOM killer
> will send a SIGTERM, do you see that in strace?
I have this on OSX, which I guess doesn't have an OOM killer;
anyway, no, the processes just hang, and no,
nothing gets killed.
OSX shoul
If its really because you are running out of memory then the OOM killer
will send a SIGTERM, do you see that in strace?
The doc builder respects the usual SAGE_NUM_THREADS environment variable.
On Wednesday, May 28, 2014 11:27:41 PM UTC+1, Dima Pasechnik wrote:
>
> On 2014-05-28, Dima Pasechnik
On 2014-05-28, Dima Pasechnik wrote:
> On 2014-05-28, Volker Braun wrote:
>> This is multiprocessing.pool waiting for children to finish.
> To finish, or to start?
>
>> The question is what happened to the children.
> They are seemingly alive, but don't do anything at all, according to
> dtruss
On 2014-05-28, Volker Braun wrote:
> This is multiprocessing.pool waiting for children to finish.
To finish, or to start?
> The question is what happened to the children.
They are seemingly alive, but don't do anything at all, according to
dtruss.
The main process occupies over 100Mb of RAM,
Ralf Stephan wrote:
How do you exclude out of memory?
Having more than 16 GB of free memory available?
-leif
On SMC for 12 hours now, there are no
more doc builds being stuck, but also no more out-of-memory which I got
yesterday.
On Wed, May 28, 2014 at 5:56 PM, Volker Braun mailto:vbraun
How do you exclude out of memory? On SMC for 12 hours now, there are no
more doc builds being stuck, but also no more out-of-memory which I got
yesterday.
On Wed, May 28, 2014 at 5:56 PM, Volker Braun wrote:
> This is multiprocessing.pool waiting for children to finish. The question
> is what h
This is multiprocessing.pool waiting for children to finish. The question
is what happened to the children.
--
You received this message because you are subscribed to the Google Groups
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to sage-d
On 2014-05-28, Travis Scrimshaw wrote:
> I was able to get it to work again (at least with 6.3.beta1) by running
> "make distclean && make doc-clean && make".
yes, I mentioned that "make distclean..." does fix things,
but this is not what one wants to do, IMHO, unless you work on a
100-core mons
On 2014-05-28, Volker Braun wrote:
> On Wednesday, May 28, 2014 1:28:58 PM UTC+1, leif wrote:
>>
>> > Strace the hanging process?
>>
>> Already did that; the problem is that it's busy waiting (i.e., polling).
>> Each time I kill the process it is waiting for [...]
>
>
> So whats that process d
I was able to get it to work again (at least with 6.3.beta1) by running
"make distclean && make doc-clean && make".
Best,
Travis
On Wednesday, May 28, 2014 6:48:13 AM UTC-7, leif wrote:
>
> Volker Braun wrote:
> > On Wednesday, May 28, 2014 1:28:58 PM UTC+1, leif wrote:
> >
> > > Strace
Volker Braun wrote:
On Wednesday, May 28, 2014 1:28:58 PM UTC+1, leif wrote:
> Strace the hanging process?
Already did that; the problem is that it's busy waiting (i.e.,
polling).
Each time I kill the process it is waiting for [...]
So whats that process doing?
As mentio
On Wednesday, May 28, 2014 1:28:58 PM UTC+1, leif wrote:
>
> > Strace the hanging process?
>
> Already did that; the problem is that it's busy waiting (i.e., polling).
> Each time I kill the process it is waiting for [...]
So whats that process doing? "strace -ff" and then sort out which
sub
On Wednesday, 28 May 2014 13:56:16 UTC+1, leif wrote:
>
> leif wrote:
> > Volker Braun wrote:
> >> Is there an error further up in the log?
> >
> > Only the usual GAP foo ("cannot extend the workspace further" or
> > something like that).
> >
> >> Strace the hanging process?
> >
> > Alre
leif wrote:
Volker Braun wrote:
Is there an error further up in the log?
Only the usual GAP foo ("cannot extend the workspace further" or
something like that).
Strace the hanging process?
Already did that; the problem is that it's busy waiting (i.e., polling).
Each time I kill the proces
Volker Braun wrote:
Is there an error further up in the log?
Only the usual GAP foo ("cannot extend the workspace further" or
something like that).
Strace the hanging process?
Already did that; the problem is that it's busy waiting (i.e., polling).
Each time I kill the process it is wai
Volker Braun wrote:
Is there an error further up in the log? Strace the hanging proces
Sometimes you need to run "sage -sync-build".
On Tuesday, May 27, 2014 11:22:06 PM UTC+1, Dima Pasechnik wrote:
Is there anything short of "make distclean && make" to fix the stuck
docbuild:
Is there an error further up in the log? Strace the hanging process?
Sometimes you need to run "sage -sync-build".
On Tuesday, May 27, 2014 11:22:06 PM UTC+1, Dima Pasechnik wrote:
>
> Is there anything short of "make distclean && make" to fix the stuck
> docbuild:
>
> By "stuck" I mean tha
It happens here on SageMathCloud which is 3.13.0-24-generic #46-Ubuntu
x86_64
but not OpenSuSE-12.3
on
AMD, at least so far.
I am not sure if it's connected to Out of
Memory errors that I'm seeing too recently on SMC.
On Wed, May 28, 2014 at 11:09 AM, Dima Pasechnik wrote:
> On Wednesday,
On Wednesday, 28 May 2014 01:12:01 UTC+1, leif wrote:
>
> Dima Pasechnik wrote:
> > Is there anything short of "make distclean && make" to fix the stuck
> > docbuild:
> >
> > By "stuck" I mean that I run make and
> >
> > the output (and logs/dochtml.log) getting stuck at
> >
> > [histo
Dima Pasechnik wrote:
Is there anything short of "make distclean && make" to fix the stuck
docbuild:
By "stuck" I mean that I run make and
the output (and logs/dochtml.log) getting stuck at
[history_a] no targets are out of date.
[structure] loading cross citations... looking for now-outda
36 matches
Mail list logo
