Problem with complex decay chains (Clustering failed in cluster.f)

Bug #1406214 reported by Josh Bendavid
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
MadGraph5_aMC@NLO
Fix Released
Undecided
Olivier Mattelaer

Bug Description

When trying to produce a gridpack for a somewhat complex LO process with complex decay chain, some of the jobs fail with

 cluster.f: Error. Invalid combination.
 Error: Clustering failed in cluster.f.

Reproducible with below example (picked out one specific process*decay to reproduce faster):

import model sm-no_b_mass
define p = p b b~
define j = j b b~
generate g g > t t~ g g g, (t > w+ b, w+ > u d~), (t~ > w- b~, w- > u~ d)

log from one of the failing jobs attached

Revision history for this message
Josh Bendavid (joshbendavid) wrote :
Revision history for this message
Olivier Mattelaer (olivier-mattelaer) wrote :

Hi Josh,

Do you have the banner of that run?
Looks like this depends on the run_card parameter since I do not reproduce this behavior with the default run_card.

Cheers,

Olivier

Revision history for this message
Josh Bendavid (joshbendavid) wrote : Re: [Bug 1406214] Re: Problem with complex decay chains (Clustering failed in cluster.f)

Hmm maybe not the exact one still. Can try to recreate it. Most relevant
change from default would probably be ickkw =1

Sent from my Android.
On 7 Feb 2015 14:30, "Olivier Mattelaer" <email address hidden>
wrote:

> Hi Josh,
>
> Do you have the banner of that run?
> Looks like this depends on the run_card parameter since I do not reproduce
> this behavior with the default run_card.
>
> Cheers,
>
> Olivier
>
> --
> You received this bug notification because you are subscribed to the bug
> report.
> https://bugs.launchpad.net/bugs/1406214
>
> Title:
> Problem with complex decay chains (Clustering failed in cluster.f)
>
> Status in MadGraph5_aMC@NLO Generator:
> New
>
> Bug description:
> When trying to produce a gridpack for a somewhat complex LO process
> with complex decay chain, some of the jobs fail with
>
> cluster.f: Error. Invalid combination.
> Error: Clustering failed in cluster.f.
>
>
> Reproducible with below example (picked out one specific process*decay
> to reproduce faster):
>
> import model sm-no_b_mass
> define p = p b b~
> define j = j b b~
> generate g g > t t~ g g g, (t > w+ b, w+ > u d~), (t~ > w- b~, w- > u~ d)
>
>
> log from one of the failing jobs attached
>
> To manage notifications about this bug go to:
> https://bugs.launchpad.net/mg5amcnlo/+bug/1406214/+subscriptions
>

Revision history for this message
Josh Bendavid (joshbendavid) wrote :

Ok, found the run card associated with the above test case. (attached)

Revision history for this message
Olivier Mattelaer (olivier-mattelaer) wrote :

Thanks, I’m relaunching the run to see if it pass or not.

Cheers,

Olivier
On 07 Feb 2015, at 13:47, Josh Bendavid <email address hidden> wrote:

> Ok, found the run card associated with the above test case. (attached)
>
> ** Attachment added: "ttdecay_run_card.dat"
> https://bugs.launchpad.net/mg5amcnlo/+bug/1406214/+attachment/4314222/+files/ttdecay_run_card.dat
>
> --
> You received this bug notification because you are subscribed to
> MadGraph5_aMC@NLO.
> https://bugs.launchpad.net/bugs/1406214
>
> Title:
> Problem with complex decay chains (Clustering failed in cluster.f)
>
> Status in MadGraph5_aMC@NLO Generator:
> New
>
> Bug description:
> When trying to produce a gridpack for a somewhat complex LO process
> with complex decay chain, some of the jobs fail with
>
> cluster.f: Error. Invalid combination.
> Error: Clustering failed in cluster.f.
>
>
> Reproducible with below example (picked out one specific process*decay to reproduce faster):
>
> import model sm-no_b_mass
> define p = p b b~
> define j = j b b~
> generate g g > t t~ g g g, (t > w+ b, w+ > u d~), (t~ > w- b~, w- > u~ d)
>
>
> log from one of the failing jobs attached
>
> To manage notifications about this bug go to:
> https://bugs.launchpad.net/mg5amcnlo/+bug/1406214/+subscriptions

Changed in mg5amcnlo:
assignee: nobody → Olivier Mattelaer (olivier-mattelaer)
Revision history for this message
Olivier Mattelaer (olivier-mattelaer) wrote :

Hi Josh,

The run was working as well...
can you track in which directory/channel that errror was raised?
I can try to run only that channel with a lot of event, and will at some point be able to reproduce the error I guess.

Cheers,

Olivier

Revision history for this message
Olivier Mattelaer (olivier-mattelaer) wrote :

Hi Josh,

Another report concerning the same error message was reported last week.
That one I was able to reproduce and this is the associate fix:

=== modified file 'Template/LO/SubProcesses/reweight.f'
--- Template/LO/SubProcesses/reweight.f 2015-02-24 00:16:04 +0000
+++ Template/LO/SubProcesses/reweight.f 2015-03-23 05:35:33 +0000
@@ -729,7 +729,7 @@
            endif
            if(ipart(2,imocl(n)).gt.2)then ! ipart(1) set and not IS line
 c The ishft gives the FS particle corresponding to imocl
- if(ipdgcl(ishft(1,ipart(2,imocl(n))-1),igraphs(1),iproc).ne.21.and.
+ if(ipdgcl(ishft(1,ipart(2,imocl(n))-1),igraphs(1),iproc).ne.21.or.
      $ ipdgcl(imocl(n),igraphs(1),iproc).ne.21) then
 c The second condition is to prevent the case of ggh where the gluon split in quark later.
 c The first quark is already remove so we shouldn't remove this one.

Cheers,

Olivier

Changed in mg5amcnlo:
status: New → Fix Committed
Changed in mg5amcnlo:
status: Fix Committed → Fix Released
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.