Opened 10 years ago
Closed 10 years ago
#606 closed defect (fixed)
Numerical noise in shower output
| Reported by: | kilian | Owned by: | kilian |
|---|---|---|---|
| Priority: | P0 | Milestone: | v2.2.0 |
| Component: | core | Version: | 2.1.1 |
| Severity: | blocker | Keywords: | |
| Cc: |
Description
The unit test parton_shower_1 fails on Jenkins because the output of nagfor and gfortran differ. (The quad mode is also different.) There are too many almost-zero numbers, so use pacify?
This is of course not a new issue, but it breaks any functional tests.
Change History (3)
comment:1 Changed 10 years ago by
| Priority: | P2 → P0 |
|---|---|
| Severity: | normal → blocker |
comment:2 Changed 10 years ago by
| Status: | new → assigned |
|---|
It's not a disaster, we just didn't check the shower output at all for numerical stability. (Regarding the 2.2 tests, we may use just a summary, e.g., the number of particles.) Anyway, I don't want to delve into that code too deeply, but maybe there is a simple solution. Let's see ...
comment:3 Changed 10 years ago by
| Resolution: | → fixed |
|---|---|
| Status: | assigned → closed |
Should be fixed in r5333. I took the opportunity of making the ascii and debug event formats somewhat more useful.
Note: See
TracTickets for help on using
tickets.
This is a desaster ... make extra-distchecks fails now since several days for several reasons... we will have to look into this.