- You can get the source in the following ways, where yourdirectoryforwf is a directory of you choice where you want to place them :
- git clone git@git.epn-campus.eu:wavefrontid01 yourdirectoryforwf
if you have access to the repository
- git clone ~mirone/src/wavefrontid01 yourdirectoryforwf
if you have access to the above directory
- go now to yourdirectoryforwf
you need to be on a computer with nvidia GPU and cuda installed. At ESRF the transition to Debian is not yet fully accomplished so you will probably land on an very old installation where the system provided python is very old and has not even numpy. Developpment Environment : To use a proper developpment environment do the following :
- source /scisoft/ESRF_sw/opteron2/set_environment
- Question N1:
how to land on a gpu workstation?
- Answer :
oarsub -p”gpu=’YES’ ” -l”cpu=1,walltime=0:30:00” -I
This ask for half an hour reserving one whole cpu on a machine with gpu. it is important to ask a whole cpu because gpu are associated to CPU. If you use just one core on a cpu, another user might have another core on the same cpu and you would finish being concurrent on the same gpu.
Do not ask to much more than you need in time. Batch submission are also possible which exit when the execution has finished
if you are using your own pc you are already there and dont need the oar call.
go to yourdirectoryforwf
source compila
- and you are ready to run
the default compila script, put the executable programs (actually scripts) into
yourdirectoryforwf/dummy/scripts/You will prepend the script name with the above path when you execute the different codes, unless differently installed.
For example at ESRF you can find already the installed scripts into
/scisoft/ESRF_sw/opteron2/PACKAGES/wf_000/scriptswhere 000 is the incremental version number (000 for the first one )
COMPILING THE DOCUMENTATION
This works probably only on a recent distribution. Dont try it on ESRF machines, you need sphinx installed :
go to yourdirectoryforwf/doc
make html make latexpdf