whizard is hosted by Hepforge, IPPP Durham

Opened 9 years ago

Closed 9 years ago

#708 closed defect (fixed)

POWHEG test fails with NAG

Reported by: Juergen Reuter Owned by: kilian
Priority: P3 Milestone: v2.2.6
Component: core Version: 2.2.5
Severity: normal Keywords:
Cc:

Description

This is the backtrace:

Runtime Error: powheg.f90, line 849: ALLOCATABLE variable is not currently allocated
Program terminated by fatal error
powheg.f90, line 849: Error occurred in POWHEG:POWHEG_SETUP_SUDAKOVS
powheg.f90, line 814: Called by POWHEG:POWHEG_CONNECT
powheg.f90, line 1493: Called by POWHEG:POWHEG_1
unit_tests.f90, line 187: Called by UNIT_TESTS:TEST
powheg.f90, line 1441: Called by POWHEG:POWHEG_TEST
whizard.f90, line 738: Called by WHIZARD:WHIZARD_CHECK
main.f90, line 357: Called by MAIN
Running test: powheg_1

Change History (2)

comment:1 Changed 9 years ago by Bijan Chokoufe Nejad

Well.. In 849 we call

       call powheg%sudakov(alr)%s%init (...)

and it gets allocated in the same subroutine (powheg_setup_sudakovs):

    allocate (powheg%sudakov (powheg%process_deps%n_alr))
    do alr = 1, powheg%process_deps%n_alr
       ...
       allocate (sudakov_simple_fsr_t :: powheg%sudakov(alr)%s)

comment:2 Changed 9 years ago by Juergen Reuter

Resolution: fixed
Status: newclosed

Two issues: powheg%rng was not allocated, that was the complaint from the NAG compiler. And the printout then generated NaNs? because the real components of powheg_t were not initialized. Resolved in r6629.

Note: See TracTickets for help on using tickets.