Hello! I am having problems plotting the histograms of my monte carlo
runs. Sometimes (I suspect due to license timeouts) the simulation of
some points in the run "kind of" fails (i.e., they are not flagged as
failed, but all their associated evaluated/measured quantities are
zero!). These points lead to wrong histograms, since the zero values are
taken into account for the plotting.
If, however, after the whole
monte carlo run is over I use the option "[right-click] Troubleshoot
point" on these failed points, they all run fine. The problem is that the
troubleshot results for these points are not integrated with the rest of
the monte carlo run, so I still get ill histograms. So my question is,
how can I tell adexl to merge the troubleshot points' results with the
rest of the original results, so as to replace the old zero-valued
entries?Thanks in advance for any help/ideas/comments.Cheers,Jorge.P.S.
I naively attempted to fool adexl by manually copying the troubleshot
results into the old point's data directory, but it didn't work; I guess
I'm still lacking a way to make adexl reload the data from the disk...
how could I force this? (yes, I tried closing everything and reopening,
but it doesn't seem to read the overwritten data!).
I don't think this is possible currently (at least not without a lot of hacking). I suggest you contact customer support - it seems a reasonable request to be able to do this.
In reply to Andrew Beckett:
CCR 1183106 has been filed for this. Everyone who also needs this feature should ask Cadence support to file a duplicate of this CCR on his behalf. The more people ask for this feature, the more likely it is to be implemented soon.
Thanks for your replies, Andrew & Frank. Unfortunately, I believe Europractice academic members don't have the possibility to contact Cadence customer support... do you know of any other means to channel such support requests for indirect Cadence users like us? Cheers,Jorge.
In reply to spectrallypure:
Contact Europractice (via http://www.europractice.stfc.ac.uk/contact_us.html I think) and then they will contact us on your behalf. We work closely with the folks at Rutherford Appleton Labs who run the Europractice software service.
Oh, that's great news Andrew; I will then contact them w.r.t to this problem.Thanks again,
I see that the CCR has been rated with "Severity: Minor" by Cadence. Do you agree with this rating? What sense does the ability to troubleshoot points make if there is no way to integrate the results? This is a pretty obvious problem with no practical workaround, so I am rather surprised about this rating. One might even argue that the correct operation would be to automatically integrate the results from the troubleshoot points and rate the current behavior as a bug.
Jorge (and others needing this feature): If you want to get this fixed soon, you absolutely need to have your own CCRs filed for this issue. Hopefully, this will cause Cadence to give a higher priority to this problem.
In reply to Frank Wiedmann:
The CCR hasn't actually been assessed to have S2 (minor) severity. What's happened is that it has been duplicated with earlier requests for this functionality. In fact it's now a member of a set of 4 duplicate CCRs - two of which are for merging the troubleshoot point results with the run they originated from, but the master is for a general request to be able to merge the results from any two runs (for example, you might have run with one sweep - and then run with another couple of points, but want to see the complete set merged together). Since the results for troubleshoot point are a different history point, effectively it's covered by the more general request.
However, the original requester for the general result merging filed it as an S2, and R&D at the time accepted that assessment of the severity.
Typically as duplicates come along, they get duplicated but nobody sits and reassesses the master's severity as the duplication is done. Instead, this happens during one of the periodic planning reviews as work is decided upon for an upcoming release. Then the severity and status etc will get updated accordingly. It just means it's done in the light of how many other issues with more (or less) impact have come up in the meantime - rather than just an impulse response on the day of a CCR being filed.
Given that 3 of the 4 duplicates were filed as S1, I've asked R&D/PE to update the master's reviewed severity to be S1 rather than S2 (and so you'd then see that severity on your CCR too - because the master data is visible on all duplicates).
The good news is that there are now already 4 in the set, so that's a good start! However, I should point out that implementing a result merging capability is not as trivial as it sounds - because there's the vital task needed of ensuring consistency in the results (less of a problem with troubleshoot point than the more general merging case).
Thanks a lot for the clarification. That sounds much better. It's a bit unfortunate that nothing of all this is visible for the customer on the support website.
The severity has now been updated, as requested.
Just to let you know: I currently still see CCR 1183106 listed with "Severity: Minor" on support.cadence.com. However, now that I know all the background, this change from S2 to S1 is not really very important for me anymore.
It's definitely been changed. Maybe support.cadence.com is showing the "reported" severity of the master CCR rather than the Cadence-assigned severity (although I'm not sure how) - or maybe it's just a data synchronization issue:
Title: Ability to merge two run results
State: Inactive Inactive Reason: BK-Backlog
Req Type: E)nhancement Rep Req Type: E)nhancement
Severity: 1 Rep Severity: 2
Urgency: D Rep Urgency: D
Seems to have been a data synchronization issue: now I also see "Severity: Important" for this CCR.