Segfaults for specific order of insert_elements commands

Moderators: michael_borland, soliday

Post Reply
nkuklev
Posts: 8
Joined: 13 Aug 2019, 09:08
Location: University of Chicago

Segfaults for specific order of insert_elements commands

Post by nkuklev » 29 Jan 2020, 09:51

Hi Michael and all,

I seem to have a weird issue with IBSCATTER, WATCH elements and &insert_elements command. Namely, depending on where in task file I try to insert these, it either works normally, messes up other computations (like moments), or straight up segfaults. I suspect it has something to do with commands requiring previous output to be of same size/element count, but the specific combinations that break are weird.

I have attached sample files, where I try to generate a matched distribution and track it after inserting the elements. Various orderings and results are commented.

In general, is the correct procedure to always insert/alter everything right after run_setup and no longer modify lattice afterwards?

-Nikita
Attachments
v83_1_1_64_32_16_16.lte
(24.31 KiB) Downloaded 177 times
iota_track.ele
(1.47 KiB) Downloaded 180 times

michael_borland
Posts: 1927
Joined: 19 May 2008, 09:33
Location: Argonne National Laboratory
Contact:

Re: Segfaults for specific order of insert_elements commands

Post by michael_borland » 07 Feb 2020, 17:54

Nikita,

The correct procedure is indeed to perform insert_elements right after the run_setup command. I'll add some code to enforce that.

Thanks for reporting the bug.

--Michael

Post Reply