VMD-L Mailing List

From: Rich Cook (rcook_at_llnl.gov)
Date: Wed Jan 20 2010 - 17:38:13 CST

Sadly, changing versions had zero effect. Same crash. :-(

On Jan 14, 2010, at 9:27 PM, John Stone wrote:

>
> Rich,
> You might try using the same version of Tcl I'm using and
> see if that has any impact. I had no trouble once I used the
> flags I sent along to you. Either way, I do strongly suggest
> reporting the problem to the Tcl developers on their bug tracking
> site. They are usually pretty responsive, but I don't know how much
> testing they do on AIX or if they've even got regular access to an
> AIX host machine.
>
> Cheers,
> John
>
> On Thu, Jan 14, 2010 at 05:03:21PM -0800, Rich Cook wrote:
>> Well, that's interesting. Your flags do not help, but I looked at
>> that thread and the original poster mentioned that make test failed.
>> I tried to do a make test in the Tcl source directory and it fails
>> exactly as VMD does. The problem is with Tcl therefore. Sigh.
>> I guess I'm going to have to figure out how to make Tcl work on AIX
>> here. This is the joy of my job! :-)
>>
>> On Jan 14, 2010, at 2:44 PM, John Stone wrote:
>>
>>>
>>> Rich,
>>> My AIX builds are based on AIX 6.1 on the NCSA "BluePrint" Blue
>>> Waters
>>> software testbed machine here at Illinois. I'm using Tcl 8.5.6
>>> and I
>>> compiled everything from source myself. When I compiled Tcl from
>>> source, I had to workaround a bug in the Tcl stack space checking
>>> that causes problems on AIX, documented here:
>>> http://**old.nabble.com/--tcl-Bugs-1889719---Stack-space-check-
>>> fails-
>>> on-AIX-td18235084.html
>>>
>>> For both VMD and NAMD, we compiled Tcl on this AIX 6.x box
>>> using configure flags (approximately) like these:
>>>
>>> define CFLAGS="-DTCL_NO_STACK_CHECK=1" at configure time, or edit
>>> the
>>> makefile to add that definition.
>>>
>>> env CC=xlc CFLAGS="-DTCL_NO_STACK_CHECK=1" --prefix=$HOME/tcl85
>>> --disable-threads --disable-shared
>>>
>>> I don't know if using xlc_r matters if you're just linking the
>>> library to
>>> code that could use pthreadds.
>>>
>>> More info on AIX stacks: https://**computing.llnl.gov/mpi/
>>> aix_stack.html
>>>
>>> You may want "-bmaxdata:0x80000000" at link time.
>>>
>>> Cheers,
>>> John Stone
>>> vmd_at_ks.uiuc.edu
>>>
>>> On Thu, Jan 14, 2010 at 01:02:33PM -0800, Rich Cook wrote:
>>>> We can also turn this around and let me ask you, what version of
>>>> Tcl/
>>>> Tk did you build against? I don't know what difference the AIX
>>>> version would make.
>>>>
>>>> On Jan 11, 2010, at 5:24 PM, Rich Cook wrote:
>>>>
>>>>> Hello,
>>>>> I have been on the injured list for a bit and am just getting back
>>>>> to you. It would be cool to meet you sometime if you ever do drop
>>>>> by.
>>>>> Here is I think the info you need. I have forgotton how to get
>>>>> the
>>>>> AIX version information.
>>>>>
>>>>> rcook_at_up041 (src): uname -a
>>>>> AIX up041 3 5 00C5D6DD4C00
>>>>> It's AIX 5.something
>>>>>
>>>>> Tcl and Tk are both 8.5.8
>>>>>
>>>>> On Dec 30, 2009, at 11:03 AM, John Stone wrote:
>>>>>
>>>>>>
>>>>>> Hi Rich,
>>>>>> I just got back from your neck of the woods. I passed
>>>>>> through Livermore twice last week while I was out in California
>>>>>> visiting relatives. Someday I'll have to make a point to visit
>>>>>> the lab and some of the people I know there when I'm out that
>>>>>> way.
>>>>>>
>>>>>> Which version of AIX is your machine using? My builds on an AIX
>>>>>> 6.x
>>>>>> test machine at NCSA didn't have any of these problems. I
>>>>>> compiled
>>>>>> the Tcl/Tk libraries for myself, but I didn't use any
>>>>>> particularly
>>>>>> unusual build flags for those compiles. In the past with certain
>>>>>> versions
>>>>>> of AIX and xlc I have had to do things like override the
>>>>>> compilation flags
>>>>>> to reduce optimization in order to get a stable build, but in my
>>>>>> recent
>>>>>> builds on the AIX 6.x machine at NCSA I didn't need to do this.
>>>>>> If you can tell me which exact Tcl/Tk minor version you're using,
>>>>>> I can try doing a build of the same rev and see if this is a
>>>>>> problem
>>>>>> with Tcl itself or an interaction with VMD and Tcl.
>>>>>>
>>>>>> Cheers,
>>>>>> John Stone
>>>>>> vmd_at_ks.uiuc.edu
>>>>>>
>>>>>> On Thu, Dec 17, 2009 at 06:08:08PM -0800, Rich Cook wrote:
>>>>>>> Yes, I have already done that and can give you any info you
>>>>>>> need.
>>>>>>>
>>>>>>> On Dec 17, 2009, at 5:52 PM, John Stone wrote:
>>>>>>>
>>>>>>>>
>>>>>>>> Rich,
>>>>>>>> I thought it was crashing in a Tk call, but I don't have your
>>>>>>>> traceback in front of me, so that was my vague recollection.
>>>>>>>> If it
>>>>>>>> is dying in Tcl itself, then we're going to have to figure out
>>>>>>>> what
>>>>>>>> is actually going on there. Can you recompile your Tcl/Tk
>>>>>>>> library
>>>>>>>> with
>>>>>>>> debugging symbols enabled (i.e. by adding -g to the compilation
>>>>>>>> flags)?
>>>>>>>>
>>>>>>>> Cheers,
>>>>>>>> John
>>>>>>>>
>>>>>>>> On Thu, Dec 17, 2009 at 05:35:12PM -0800, Rich Cook wrote:
>>>>>>>>> It crashes when I disable Tk in the same place as before. I'm
>>>>>>>>> confused, because you are telling me to disable Tk but it is
>>>>>>>>> crashing
>>>>>>>>> in a Tcl call. Is that intentional? I'm recompiling without
>>>>>>>>> Tcl
>>>>>>>>> and
>>>>>>>>> without Tk just to see what happens...
>>>>>>>>>
>>>>>>>>
>>>>>>>> --
>>>>>>>> NIH Resource for Macromolecular Modeling and Bioinformatics
>>>>>>>> Beckman Institute for Advanced Science and Technology
>>>>>>>> University of Illinois, 405 N. Mathews Ave, Urbana, IL 61801
>>>>>>>> Email: johns_at_ks.uiuc.edu Phone: 217-244-3349
>>>>>>>> WWW: http://****www.****ks.uiuc.edu/~johns/ Fax:
>>>>>>>> 217-244-6078
>>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> /* A function that takes a single integer argument and returns
>>>>>>> a pointer to a function that takes two integer arguments and
>>>>>>> returns a floating-point number. */
>>>>>>> float (*func2(int a))(int, int);
>>>>>>> Rich Cook
>>>>>>> rcook_at_llnl.gov
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>
>>>>>> --
>>>>>> NIH Resource for Macromolecular Modeling and Bioinformatics
>>>>>> Beckman Institute for Advanced Science and Technology
>>>>>> University of Illinois, 405 N. Mathews Ave, Urbana, IL 61801
>>>>>> Email: johns_at_ks.uiuc.edu Phone: 217-244-3349
>>>>>> WWW: http://***www.***ks.uiuc.edu/~johns/ Fax: 217-244-6078
>>>>>>
>>>>>
>>>>>
>>>>> /* A function that takes a single integer argument and returns
>>>>> a pointer to a function that takes two integer arguments and
>>>>> returns a floating-point number. */
>>>>> float (*func2(int a))(int, int);
>>>>> Rich Cook
>>>>> rcook_at_llnl.gov
>>>>>
>>>>>
>>>>>
>>>>>
>>>>
>>>>
>>>> /* A function that takes a single integer argument and returns
>>>> a pointer to a function that takes two integer arguments and
>>>> returns a floating-point number. */
>>>> float (*func2(int a))(int, int);
>>>> Rich Cook
>>>> rcook_at_llnl.gov
>>>>
>>>>
>>>>
>>>
>>> --
>>> NIH Resource for Macromolecular Modeling and Bioinformatics
>>> Beckman Institute for Advanced Science and Technology
>>> University of Illinois, 405 N. Mathews Ave, Urbana, IL 61801
>>> Email: johns_at_ks.uiuc.edu Phone: 217-244-3349
>>> WWW: http://**www.**ks.uiuc.edu/~johns/ Fax: 217-244-6078
>>>
>>
>>
>> /* A function that takes a single integer argument and returns
>> a pointer to a function that takes two integer arguments and
>> returns a floating-point number. */
>> float (*func2(int a))(int, int);
>> Rich Cook
>> rcook_at_llnl.gov
>>
>>
>>
>
> --
> NIH Resource for Macromolecular Modeling and Bioinformatics
> Beckman Institute for Advanced Science and Technology
> University of Illinois, 405 N. Mathews Ave, Urbana, IL 61801
> Email: johns_at_ks.uiuc.edu Phone: 217-244-3349
> WWW: http://*www.*ks.uiuc.edu/~johns/ Fax: 217-244-6078
>

/* A function that takes a single integer argument and returns
    a pointer to a function that takes two integer arguments and
    returns a floating-point number. */
float (*func2(int a))(int, int);
Rich Cook
rcook_at_llnl.gov