search

lwang-astro / PeTar

PeTar is a high-performance N-body code for modelling the evolution of star clusters and tidal streams, including the effect of galactic potential, dynamics of binary and hierarchical system, single and binary stellar evolution.
MIT License
71 stars 19 forks source link

BSE on M2 chip #39

Closed christianboily closed 1 year ago

christianboily commented 1 year ago

Hi Long Wang,

I have success with PeTar on an M2 processor (+mpi, galpy, amp) but without the BSE module which breaks down at compilation :

mpicxx-openmpi-gcc12 -D BSEBBF -D BSE_BASE -fpermissive -O3 -Wall -std=c++17 -fopenmp bse_test.cxx -o petar.bse -L./ -lbse -lgfortran In file included from /opt/local/include/gcc12/c++/cassert:44, from bse_interface.h:4, from bse_test.cxx:7: bse_interface.h: In member function 'bool BSEManager::checkParams()': bse_interface.h:1088:16: error: 'trackmode' was not declared in this scope 1088 | assert(trackmode>0); | ^~~~~ make[1]: [petar.bse] Error 1 make: [bse-interface/libbse.a] Error 2

--> this seems like a brutal mistake, a parameter that fails to initialise : never had that before and am wondering --> if the fault is with gcc12 ? That's the default version I can install with MacPorts (to have the mpi wrapper).

Any idea what may be causing this ? I have cloned PeTar (as well as FDPS and SDAR) to have the latest versions of the stable branches in each case.

Cheers, chris b.

christianboily commented 1 year ago

Re- in fact, the configuration with interrupt set to bseEmp (not bse) works fine : the track mode fnc is defined and the code compiles ok.

lwang-astro commented 1 year ago

Hi chris,

Thank you for pointing out this. It is a bug introduced from last update of BSEEMP. I have fixed the issue. You can now compile BSE.

christianboily commented 1 year ago

Dear Long Wang, thank you! We are interested in extreme metal poor stars here, too, cf. The Pristine survey, and it's great to have both libraries for comparison : keep up the good work!

Cheers, Chris b.

----- Mail d’origine ----- De: lwang-astro @.> À: lwang-astro/PeTar @.> Cc: christian boily @.>, Author @.> Envoyé: Thu, 01 Jun 2023 09:05:06 +0200 (CEST) Objet: Re: [lwang-astro/PeTar] BSE on M2 chip (Issue #39)

Hi chris,

Thank you for pointing out this. It is a bug introduced from last update of BSEEMP. I have fixed the issue. You can now compile BSE.

-- Reply to this email directly or view it on GitHub: https://github.com/lwang-astro/PeTar/issues/39#issuecomment-1571476820 You are receiving this because you authored the thread.

Message ID: @.***>