- 23 Feb, 2018 2 commits
-
-
Peter W. Draper authored
-
James Willis authored
-
- 22 Feb, 2018 1 commit
-
-
Peter W. Draper authored
Note also needs --enable-compiler-warnings=yes and --disable-vec and isn't known to run!
-
- 21 Feb, 2018 2 commits
-
-
Matthieu Schaller authored
-
Matthieu Schaller authored
-
- 15 Jan, 2018 2 commits
-
-
Matthieu Schaller authored
-
Matthieu Schaller authored
-
- 09 Jan, 2018 1 commit
-
-
James Willis authored
Added a branching function for DOSELF_SUBSET. To call vectorised version when vectorisation is enabled.
-
- 05 Jan, 2018 1 commit
-
-
Matthieu Schaller authored
Re-introduce the 'end_force' function as a task in order for the cooling function to be called at the right place in the time-step.
-
- 03 Jan, 2018 1 commit
-
-
James Willis authored
-
- 20 Dec, 2017 1 commit
-
-
James Willis authored
Added a branching function for dopair_subset, to switch between naive, serial and vectorised versions.
-
- 19 Dec, 2017 1 commit
-
-
Matthieu Schaller authored
Removed the now obsolete checks that all correct tasks are active. Will need to be re-done correctly.
-
- 18 Dec, 2017 2 commits
-
-
Matthieu Schaller authored
In kick1 after all is done, update the part's acceleration to include the gpart acceleration. This allows the hydro drift to use the correct accelerations for hydro predictions.
-
Matthieu Schaller authored
-
- 13 Dec, 2017 1 commit
-
-
James Willis authored
-
- 11 Dec, 2017 2 commits
-
-
Matthieu Schaller authored
-
Matthieu Schaller authored
-
- 10 Dec, 2017 2 commits
-
-
Matthieu Schaller authored
-
Matthieu Schaller authored
Separate the time-step information for hydro and gravity tasks into two different set of variables for each cell.
-
- 08 Dec, 2017 1 commit
-
-
James Willis authored
Created a branching function doself1 and doself2 to switch between naive, serial and vectorised versions.
-
- 24 Nov, 2017 1 commit
-
-
Matthieu Schaller authored
-
- 23 Nov, 2017 3 commits
-
-
Matthieu Schaller authored
-
Matthieu Schaller authored
-
Matthieu Schaller authored
Fixed a moronic bug. Unskipping MPI sends and recvs for gravity should not be done *only* for local pairs. Obviously....
-
- 05 Nov, 2017 1 commit
-
-
Matthieu Schaller authored
-
- 29 Oct, 2017 1 commit
-
-
James Willis authored
-
- 26 Oct, 2017 1 commit
-
-
James Willis authored
-
- 09 Oct, 2017 2 commits
-
-
Matthieu Schaller authored
-
Matthieu Schaller authored
-
- 06 Oct, 2017 2 commits
-
-
James Willis authored
Created a branching function for DOPAIR2 to only call the vectorised function when vectorisation is enabled and call the serial version when corner cells interact. Moved some debug statements inside the branching functions.
-
Matthieu Schaller authored
-
- 05 Oct, 2017 1 commit
-
-
Matthieu Schaller authored
-
- 01 Oct, 2017 2 commits
-
-
Matthieu Schaller authored
All gpart now find all of their interactions following correct boundary condition check in runner_do_grav_long_range()
-
Matthieu Schaller authored
-
- 29 Sep, 2017 1 commit
-
-
Matthieu Schaller authored
-
- 28 Sep, 2017 1 commit
-
-
Matthieu Schaller authored
All-reduce all the top-level multipoles after a rebuild() to make sure we have all information in hands to construct the tasks.
-
- 27 Sep, 2017 1 commit
-
-
Matthieu Schaller authored
Also send dx_max_part and dx_max_gpart alongside ti_end_min at the end of a time-step when running with MPI.
-
- 22 Sep, 2017 1 commit
-
-
James Willis authored
-
- 12 Sep, 2017 1 commit
-
-
Matthieu Schaller authored
Force compiler to report shadowed variables. Removed all shadowed variables. Added the function scheduler_activate_send() to simplify the unskipping of MPI sends.
-
- 11 Sep, 2017 1 commit
-
-
Matthieu Schaller authored
-