Opened 3 months ago

Closed 3 months ago

Last modified 3 months ago

#2701 closed Defect (fixed)

wrong periodicity on gphi

Reported by: smasson Owned by:
Priority: low Milestone:
Component: USR Version: trunk
Severity: minor Keywords:
Cc: Branch review:
MP ready?: Task progress:

Description

Context

in tests cases (at least in BENCH), periodicity is not properly applied on gphit (no calls to lbc_lnk).

Analysis

In tke, when nn_htau = 1, htau is computed with gphit, so its periodicity will also be wrong.
htau is next used to compute pavm (avm_k in zdfphy).

  • if nn_hls = 1, there is a call to lbc_lnk which will correct avm_k periodicity
  • if nn_hls = 2, we skip this call to lbc_lnk and avm_k periodicity remains wrong.

⇒ result differs with nn_hls = 1 or 2.

Recommendation

Make sure that periodicity on gphit is properly defined.
I think it is safe to apply it on all glam/gphi arrays at least when using usrdef.

Commit History (2)

ChangesetAuthorTimeChangeLog
15056smasson2021-06-25T09:37:44+02:00

trunk: correct [15052], #2701

15052smasson2021-06-24T16:39:14+02:00

trunk: call lbc_lnk for variables defined with usr_def_*, #2701

Change History (3)

comment:1 Changed 3 months ago by smasson

In 15052:

trunk: call lbc_lnk for variables defined with usr_def_*, #2701

comment:2 Changed 3 months ago by smasson

  • Resolution set to fixed
  • Status changed from new to closed

fixed in [15052], extended to all variables defined by usr_def_hgr, usr_def_zgr, usr_def_istate

comment:3 Changed 3 months ago by smasson

In 15056:

trunk: correct [15052], #2701

Note: See TracTickets for help on using tickets.