Commits

Branch selector

 master

User selector

All users

All time

Commit History

Commits on Mar 30, 2026

• Add option to rescale restart velocities

  

  mewall

  authored and

  cnegre

  committed
  c096641

  Copy full SHA for c096641

Commits on Mar 26, 2026

• Call ppot without updating parameters in MD loop

  mewall

  committed
  c1f480e

  Copy full SHA for c1f480e

• Fix acc update in ppots

  mewall

  committed
  0bd669d

  Copy full SHA for 0bd669d

• Initialize some resized matrices

  mewall

  committed
  3fdc774

  Copy full SHA for 3fdc774

• Bug fixes in kernel and neighborlist for non-offload

  mewall

  committed
  256ae35

  Copy full SHA for 256ae35

• Continuing to debug

  mewall

  committed
  fc1b587

  Copy full SHA for fc1b587

• Debug sparse neighbor list

  mewall

  committed
  5f0064c

  Copy full SHA for 5f0064c

Commits on Mar 23, 2026

• Show pair potentials upon initialization. Minor bugfix.

  mewall

  committed
  8506b7b

  Copy full SHA for 8506b7b

• Fix bug when DoKernel= F

  mewall

  committed
  5c7c52a

  Copy full SHA for 5c7c52a

Commits on Mar 17, 2026

• Fix Ewald recip bug

  mewall

  committed
  68336da

  Copy full SHA for 68336da

Commits on Mar 16, 2026

• Add use iso_c_binding where needed

  

  mewall

  authored and

  cnegre

  committed
  b30dbbf

  Copy full SHA for b30dbbf

• Revert to non-offload musearch code

  

  mewall

  authored and

  cnegre

  committed
  3cd01d1

  Copy full SHA for 3cd01d1

• Move !$acc directive inside USE_OFFLOAD region

  

  mewall

  authored and

  cnegre

  committed
  8972061

  Copy full SHA for 8972061

• Restore prg_get_hscf for offload build

  

  mewall

  authored and

  cnegre

  committed
  8e9619e

  Copy full SHA for 8e9619e

• Revert to calling prg_get_hscf_v2 (not offloaded) in the rankN update

  

  mewall

  authored and

  cnegre

  committed
  2350c2c

  Copy full SHA for 2350c2c

• A bit more threading

  

  mewall

  authored and

  cnegre

  committed
  db77553

  Copy full SHA for db77553

• Numerically correct offloaded graph update

  

  mewall

  authored and

  cnegre

  committed
  4d2b406

  Copy full SHA for 4d2b406

• Tidy up graph update before offload work

  

  mewall

  authored and

  cnegre

  committed
  3b7526a

  Copy full SHA for 3b7526a

Commits on Mar 5, 2026

• Revert gpmdcov_part.F90 to 65d376a496 commit (no offload)

  mewall

  committed
  dbce9d9

  Copy full SHA for dbce9d9

• Added flags to select custom or random seeds for adding velocity and a flag for the maximum number of updates allowed for the graph (default 200)

  

  rcorrigan

  authored and

  mewall

  committed
  9ca081f

  Copy full SHA for 9ca081f

• Calculate egap correctly

  mewall

  committed
  44b62ea

  Copy full SHA for 44b62ea

• Add a separate AnnealSteps= option to the input deck

  mewall

  committed
  128e7ef

  Copy full SHA for 128e7ef

• More optimizations in DM build

  mewall

  committed
  bd9a638

  Copy full SHA for bd9a638

• Optimizations in deortho and DM build

  mewall

  committed
  e7ad895

  Copy full SHA for e7ad895

• Optimize mu bisection search

  mewall

  committed
  2e8e70d

  Copy full SHA for 2e8e70d

• Working MPI restart

  mewall

  committed
  791c8f0

  Copy full SHA for 791c8f0

• Working on MPI-friendly restart

  mewall

  committed
  1c0e538

  Copy full SHA for 1c0e538

Commits on Mar 3, 2026

• First offload of ppot

  mewall

  committed
  bb411b2

  Copy full SHA for bb411b2

• Prepare the ppot code for offload

  mewall

  committed
  27c2957

  Copy full SHA for 27c2957

• Tidy up the Ewald real nlist code

  mewall

  committed
  3b05e46

  Copy full SHA for 3b05e46

• Working Ewald real with persistent GPU nlist

  mewall

  committed
  9f8a208

  Copy full SHA for 9f8a208

• Preserve nlist allocation. Use existing GPU pointers in Ewald real.

  mewall

  committed
  60f0ad3

  Copy full SHA for 60f0ad3

• Handle GPU memory management of neighborlist at creation/destruction

  mewall

  committed
  41170fa

  Copy full SHA for 41170fa

Commits on Mar 1, 2026

• GPU accelerated graph update (memory intensive)

  mewall

  committed
  ce78bdb

  Copy full SHA for ce78bdb

• Add option to symmetrize the graph

  mewall

  committed
  65d376a

  Copy full SHA for 65d376a