Une maintenance aura lieu le jeudi 22 janvier. Le service sera indisponible l'après-midi à partir de 14h.

Tags

Tags give the ability to mark specific points in history as being important

v2.12
protected

0c1ec9ba · v2.12 · Oct 24, 2025
broken

04258d9a · WIP PTA: introduce task list, that should simplify pipelining for GPU · May 08, 2025
TEST_OK

06c04329 · WIP PTA: fix run without MPI (some initialization was done only for MPI...) · May 05, 2025
v2.11
protected

b9e5d28b · v2.11 · Jan 11, 2025
v2.10.1
protected

c4fba466 · v2.10.1 · Dec 06, 2024
v2.10
protected

12ae6af6 · v2.10 · Nov 18, 2024
v2.9
protected

42edbc2e · v2.9 · Apr 17, 2024
gpu-fp32

ed2bca20 · GPU: use fp32 for non-linear terms. Observed speed-up is about x1.8 · Jun 30, 2023
v2.8
protected

ed67be75 · v2.8 · Mar 21, 2023
v2.7
protected

fc0d90f8 · v2.7 · Jul 12, 2022
gpu_full_ok

3228dab1 · WIP GPU: remove some cudaDeviceSynchronize() · May 22, 2022
gpu_full_port_1gpu

24dc55c2 · WIP GPU: solve_up and solve_dn on gpu · May 18, 2022
gpu_hyb_fast

4bef16a8 · WIP GPU: small improvement in case of fixed time-step · May 17, 2022
test_variation_asyn

a62cd1a2 · WIP GPU: also start computation ASAP on GPU for B-field · May 17, 2022
gpu_hackaton_start

72f4492c · bugfix: ensure initial magnetic field has a zero toroidal component at the boundaries. · May 05, 2022
gpu_ok_compared_to_cpu_6digits

303e2a0e · fix bad bug · Feb 28, 2022
gpu_ok_poltor_layout_1

ff693f63 · WIP5: gpu explicit terms: better async + support for POLTOR_LAYOUT=1 · Feb 14, 2022
gpu_ok

7c478d73 · use a compute stream (only works with comp_stream = 0) · Feb 14, 2022
v2.6
protected

d4e44d48 · v2.6 · Dec 09, 2021
v2.5.4
protected

f757dbbf · update to v2.5.4 · Mar 03, 2021

1
2
3