WATCH element Pelevant vs Elegant behavior

Moderators: cyao, michael_borland

Post Reply
nadolski
Posts: 24
Joined: 20 Mar 2013, 05:03
Location: Synchrotron SOLEIL, Gif-sur-Yvette, France
Contact:

WATCH element Pelevant vs Elegant behavior

Post by nadolski » 08 Dec 2021, 09:09

Dear All,

I am running a very simple example with the track command.
The output is different between elegant and Pelegant.

elegant mytest.ele produce mytest.w1 as expected

mpirun -n 2 Pelegant mytest.ele does not produce any mytest.w1 file.

Could you please explain why ?
Is this a bug ?

Best regards,

Laurent.
Attachments
mytest.tgz
(2.28 KiB) Downloaded 92 times

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

Re: WATCH element Pelevant vs Elegant behavior

Post by michael_borland » 08 Dec 2021, 13:08

Laurent,

The problem is that you are asking to do parallel tracking with only 1 simulation particle. It is a bug, I think, but perhaps an understandable one.

If you increase the number of particles to at least equal the number of cores, it will be fine. Typically to get a benefit from parallel tracking you should use "many" particles per core. Otherwise, interprocess communication may negate any benefit from parallelism.

How many you need to get a benefit will depend on the particular simulation.

--Michael

nadolski
Posts: 24
Joined: 20 Mar 2013, 05:03
Location: Synchrotron SOLEIL, Gif-sur-Yvette, France
Contact:

Re: WATCH element Pelevant vs Elegant behavior

Post by nadolski » 08 Dec 2021, 16:11

It was working nicely in the past.
Thanks Michael for the quick answer.
Now I now how to modify my code.

Laurent.

Post Reply