Skip to content

forest-ahmed-body

English | Japanese

« Home · « External flow and wind engineering

Ahmed body external aerodynamics (ERCOFTAC case082; Ahmed et al. 1984) on the FOREST (block-AMR) solver, per docs/design/mpi-amr-design.md §16 decisions A2 (pooled wall-data layout) / B2-in-spirit (rebuild driven by the surface sensor) / C-alpha (Bouzidi interpolated bounce-back). Same physical case as benchmarks/ahmed_body.cpp (the dense two-level solver): U=40 m/s, nu=1.5e-5, 25 deg slant, target Cd ~ 0.285. Reference (OpenFOAM SPHERIC-training AhmedBody + ERCOFTAC).

Geometry

forest-ahmed-body geometry

Fields

The fields this computation wrote at its last output step, plotted as they are. 3D cases show the mid-plane.

forest-ahmed-body p_pa

forest-ahmed-body umag_ms

Plots

Time series and distributions of the QoI written by this computation. Where a reference value exists it is overlaid as a horizontal line.

forest-ahmed-body cd-cl

Build & run

cmake -S native -B native/build && cmake --build native/build -j
./native/build/orheo-forest-ahmed-body

Cost of the computation that made the figures on this page (Miyabi-G Grace CPU (72 cores, -stdpar=multicore)): 60.7 s wall.

Validation

Tier T2 — single-point comparison
QoI Cd
Reference ERCOFTAC case082 (Ahmed 25°)
Gate — no gate

Source & run details

View source

Item Value
Target orheo-forest-ahmed-body
Source native/benchmarks/forest_ahmed_body.cpp
Lattice —
GPU twin yes (orheo-forest-ahmed-body-gpu)
Modules ahmed_solver.hpp, forest_inflow.hpp, forest_regrid.hpp, forest_step.hpp
Gallery run 2026-07-27 @ 7628c0cf
Args 8 4000 0.05 1200 0.50 25 1 1 .
Wall time 60.66 s
CPU time 4225.62 s
Device Miyabi-G Grace CPU (72 cores, -stdpar=multicore)

Computation log

The actual output of this computation (head and tail).

=== forest Ahmed body ===
nh=8 steps=4000 ulat=0.0500 re=1.2e+03 cs=0.500 slant=25.0 band=1 max_level=1
dx0=0.03600m  domain blocks=105x14x10 (kNb=4)  tau_coarse=0.5010
=== building surface-following forest ===
forest: 16254 leaves, max_level reached=1 (requested 1)
=== building wall data (solid mask + Bouzidi qlink) ===
solid cells=44834 (of 3510864 pool cells)
point probe at world (x=0.310, z=0.234) -> probe.csv
=== time-stepping ===
step      0/4000  t_conv=0.000  Cd=18.8250  Cl=4.1482  umax=0.107 | 0.0s  0 step/s  0.0 MLUPS | Mdrift=+0.0 ppm
step   1000/4000  t_conv=0.119  Cd=1.0324  Cl=-0.5324  umax=0.090 | 14.5s  69 step/s  72.0 MLUPS | Mdrift=+0.0 ppm
step   2000/4000  t_conv=0.238  Cd=1.2686  Cl=-0.0402  umax=0.090 | 28.1s  71 step/s  74.0 MLUPS | Mdrift=+0.0 ppm
step   3000/4000  t_conv=0.357  Cd=1.2889  Cl=0.0709  umax=0.090 | 44.2s  68 step/s  70.6 MLUPS | Mdrift=+0.0 ppm
step   4000/4000  t_conv=0.476  Cd=1.2550  Cl=-0.0339  umax=0.090 | 60.2s  66 step/s  69.1 MLUPS | Mdrift=+0.0 ppm