integer overflow in relaxpol
Bug #1248047 reported by
Vittorio
This bug affects 1 person
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
ABINIT |
Fix Committed
|
Medium
|
jmbeuken |
Bug Description
In relaxpol.F90 lines 471 and 473 I see the subexpression 10**ii
I believe this produces overflows when ii is too large (ii>9).
Considering that the variable "value" is real(dp),
I suggest changing 10 into 10.0_dp to get the larger range of double precision.
This way you avoid overflow and get a precise result,
thus truly eliminating noise.
Changed in abinit: | |
importance: | Undecided → Critical |
importance: | Critical → Medium |
status: | New → Confirmed |
assignee: | nobody → jmbeuken (jean-michel-beuken) |
status: | Confirmed → In Progress |
Changed in abinit: | |
status: | In Progress → Fix Committed |
To post a comment you must log in.