Project

General

Profile

Activity

From 15/04/2014 to 14/05/2014

14/05/2014

18:00 Revision 1114 (fileas): remove print
cmercier
17:58 Revision 1113 (fileas): indentation
cmercier
17:02 Revision 1112 (fileas): Little modifications in hek_dialogue definitons (convert every fonctions which were outside the class into static methods of the class) and updates induced in the other files. Use also the new method to create links between events suggested by Ryan Timmons.
jgoujon
16:24 Revision 1111 (fileas): print
cmercier
16:16 Revision 1110 (fileas): less print
cmercier
11:37 Revision 1109 (fileas): gpu/phi selection
cmercier
11:18 Revision 1108 (fileas): octopus on phi
cmercier
11:16 Revision 1107 (fileas): ajout acc_get_device pour GPU/PHI
cmercier
11:14 Revision 1106 (fileas): test
cmercier
10:45 Revision 1105 (fileas): pour test k20
cmercier
10:39 Revision 1104 (fileas): upadte of the hek_dialogue test file
jgoujon
10:27 Revision 1103 (fileas): commentaire
cmercier

06/05/2014

17:36 Revision 1102 (fileas): Update of the way we send the results to the HEK and change the parameter probability to use the parameter importance (email of marc cheung 22_04_2014)
jgoujon

29/04/2014

16:30 Revision 1101 (fileas): temporary version commit for test on gpu
cmercier

25/04/2014

10:09 Revision 1100 (fileas): Replace unwanted tests
jgoujon

16/04/2014

15:54 Revision 1099 (fileas): modifications to send the results to the HEK. The presence of the data in the HEK are deduced from the log_hek.txt file, if an ivorn is sotred in this file, it is supposed to be already on the HEK. To transfer multiple results contained in a period, we firstly upload the filaments without the links to preceding and following ivorns. In a second time, the results are resent to the HEK with the links updated, this allows to avoid problems if the ivorns are not presents in the HEK.
jgoujon
 

Also available in: Atom