Issues

Independent clamping of face-averaged volume fractions can inflate mixture properties

#1911

· sbryngelson opened

on Sep 19, 2026

Does the viscous azimuthal term have the divergence-side 1/r in 3D cylindrical?

#1910

· sbryngelson opened

on Sep 19, 2026

Energy diffusion flux applied twice for chemistry + surface_tension without viscous

#1909

· sbryngelson opened

on Sep 19, 2026

flux_src_n energy-slot deallocation lacks the chemistry guard its allocation has

#1908

· sbryngelson opened

on Sep 19, 2026

Chemistry species entries of flux_src_n are allocated but never deallocated

#1907

· sbryngelson opened

on Sep 19, 2026

MPI_FILE_OPEN's ierr is never checked, so a failed open surfaces as an unexplained crash later

#1903

· sbryngelson opened

on Sep 18, 2026

Restarting a moving IB re-measures the centre of mass from the body at the restart attitude, so the kinematics change reference point across a restart

#1902

· sbryngelson opened

on Sep 18, 2026

s_ibm_correct_state writes q_prim inside the loop that reads q_prim at image-point stencils: read-after-write race across ghost points

#1901

· sbryngelson opened

on Sep 18, 2026

The moving-IB ghost-point budget is twice the GLOBAL total used as a per-rank bound, and is never checked: silent device memory fault

#1900

· sbryngelson opened

on Sep 18, 2026

IB force relay and lookup run on the device, and a rank that has just gained a patch reads a stale device copy (fault, or zero forces)

#1899

· sbryngelson opened

on Sep 18, 2026

A patch that leaves a rank's IB neighbourhood can never re-enter it: s_handoff_ib_ownership forwards only to immediate neighbours

#1898

· sbryngelson opened

on Sep 18, 2026

GPU fault at an IB ownership handoff: kernels launched with zero-size mapped arrays on ranks that track no patch

#1897

· sbryngelson opened

on Sep 18, 2026

Navigation Menu

All issues

Independent clamping of face-averaged volume fractions can inflate mixture properties

Does the viscous azimuthal term have the divergence-side 1/r in 3D cylindrical?

Energy diffusion flux applied twice for chemistry + surface_tension without viscous

flux_src_n energy-slot deallocation lacks the chemistry guard its allocation has

Chemistry species entries of flux_src_n are allocated but never deallocated

MPI_FILE_OPEN's ierr is never checked, so a failed open surfaces as an unexplained crash later

Restarting a moving IB re-measures the centre of mass from the body at the restart attitude, so the kinematics change reference point across a restart

s_ibm_correct_state writes q_prim inside the loop that reads q_prim at image-point stencils: read-after-write race across ghost points

The moving-IB ghost-point budget is twice the GLOBAL total used as a per-rank bound, and is never checked: silent device memory fault

IB force relay and lookup run on the device, and a rank that has just gained a patch reads a stale device copy (fault, or zero forces)

A patch that leaves a rank's IB neighbourhood can never re-enter it: s_handoff_ib_ownership forwards only to immediate neighbours

GPU fault at an IB ownership handoff: kernels launched with zero-size mapped arrays on ranks that track no patch

Search results