sPHENIX discussion of software

["Michael P. McCumber" <mccumber AT bnl.gov>]

Yep, I pulled over last night's build and that cleared the compilation problems.

Thanks!

Mike

--

Michael P. McCumber, PhD
Los Alamos National Laboratory
505-709-8161

On Fri, Jan 8, 2016 at 8:20 PM, pinkenburg <pinkenburg AT bnl.gov> wrote:

Hi Mike,

it's not a link dependence in some makefile, that's libtool. It picks up some *.la file and tries to locate the la-files mentioned in there. The reason can be an old install area, trying to link while a build has wiped out this install area.

I normally remove those la files from our installed software - they cause nothing but trouble and are not needed for gcc (and we remove them during the build which is why you don't see them in $OFFLINE_MAIN/lib). There were indeed some left in gsl (the ones you see here) which is copied to $OFFLINE_MAIN as part of the nightly build. I removed them and started a fresh build. That should flush those out.

You probably need to clear out your install area to get rid of old la files which mention those.

Chris

On 1/8/2016 9:52 PM, Michael P. McCumber wrote:

Hi Chris,

I've updated the gcc, and can run the code, but when I try to recompile the g4detectors library I get the following error:

/bin/grep: /afs/rhic.bnl.gov/sphenix/sys/x8664_sl6/new.1/lib/libgsl.la: No such file or directory

/bin/sed: can't read /afs/rhic.bnl.gov/sphenix/sys/x8664_sl6/new.1/lib/libgsl.la: No such file or directory

lib tool: link: `/afs/rhic.bnl.gov/sphenix/sys/x8664_sl6/new.1/lib/libgsl.la' is not a valid libtool archive

I expect that there is somewhere in the code base with an explicit link dependence through to afs which isn't available in remote systems instead of linking through to /opt/sphenix.

Any ideas on how to best chase that down?

Mike

--

Michael P. McCumber, PhD
Los Alamos National Laboratory
505-709-8161

On Tue, Jan 5, 2016 at 1:55 PM, Michael P. McCumber <mccumber AT bnl.gov> wrote:

No impact on tracking performance. See attachment.

--

Michael P. McCumber, PhD
Los Alamos National Laboratory
505-709-8161

On Mon, Jan 4, 2016 at 10:05 AM, Chris Pinkenburg <pinkenburg AT bnl.gov> wrote:

Hello Mike,

it is based on the redhat developer toolsets. The dev toolset does work but the binaries don't run on our farm (in my understanding). The rpms to install the tweaked rcf version are available in afs (you need a token, it's not readable without like our afs space):

/afs/rhic.bnl.gov/rcfsl/6.4/rcf-custom/gcc-4.8.2

Chris

On 01/04/2016 10:27 AM, Michael P. McCumber wrote:

Hello Chris,

For those of us---aka me---using non-RCF resources, is this just a matter of installing stock version of gcc 4.8.2? Or are the custom hacks documented for us to follow?

Mike

--

Michael P. McCumber, PhD
Los Alamos National Laboratory
505-709-8161

On Mon, Jan 4, 2016 at 6:45 AM, pinkenburg <pinkenburg AT bnl.gov> wrote:

Hi folks,

the nightly build using gcc 4.8 and Geant4 10.02 is working. I would
like to switch to this as soon as possible so we can start to use C++11.
I reran the Atlanta hcal fest single particle sims  (1d/2d spacal and
zero tilt angle neutrons), the output is under:

/gpfs02/phenix/prod/sPHENIX/gcc4.8/single_particle/(spacal1d,spacal2d,hcalstudies/0cross)/fieldmap

To analyze those files you need to use the play build

source /opt/sphenix/core/bin/sphenix_setup.csh -n play

If you want to use gcc 4.8 to compile your code you have to tweak the
path (will be part of the setup script once we switch):

set path = (/opt/gcc/4.8.2/bin $path)

But you don't have to Chris Hollowell from rcf tweaked the gcc 4.8
installation that this can run together with our standard compiler.

Please have a look at this at your earliest convenience

Chris

--
*************************************************************

Christopher H. Pinkenburg       ;    pinkenburg AT bnl.gov
                                ;    http://www.phenix.bnl.gov/~pinkenbu

Brookhaven National Laboratory  ;    phone: (631) 344-5692
Physics Department Bldg 510 C   ;    fax:   (631) 344-3253
Upton, NY 11973-5000

*************************************************************

_______________________________________________
Sphenix-software-l mailing list
Sphenix-software-l AT lists.bnl.gov
https://lists.bnl.gov/mailman/listinfo/sphenix-software-l

-- 

*********************************************************

Christopher H. Pinkenburg      ; pinkenburg AT bnl.gov

Brookhaven National Laboratory ; phone: (631) 344-5692
Physics Department Bldg 510 C  ; fax:   (631) 344-3253
Upton, NY 11973-5000

http://www.phenix.bnl.gov/~pinkenbu/

*********************************************************

-- 
*************************************************************

Christopher H. Pinkenburg	;    pinkenburg AT bnl.gov
				;    http://www.phenix.bnl.gov/~pinkenbu

Brookhaven National Laboratory	;    phone: (631) 344-5692
Physics Department Bldg 510 C	;    fax:   (631) 344-3253
Upton, NY 11973-5000

*************************************************************