Add a Warp (NVIDIA warp-lang) GPU backend for PySPH WCSPH - #435

Open

kunalpuri-prediqt wants to merge 54 commits into

kunalpuri-prediqt:blast-from-the-past

Commits on Jun 15, 2026

Commits on Jun 18, 2026

Fuse Warp continuity-density PEC equations via code generation (ADR-0003)

kunalpuri-prediqt
and
claude
committed
Grid-direct neighbor traversal for the WCSPH continuity hot path (ADR-0004)

kunalpuri-prediqt
and
claude
committed
Add BUILD.md: reproducible build + GPU compatibility guide

kunalpuri-prediqt
and
claude
committed
Migrate remaining neighbor-loop kernels onto the generator (ADR-0003 follow-up)

kunalpuri-prediqt
and
claude
committed
Grid-direct the summation Euler/KDK step paths (grid-direct everywhere)

kunalpuri-prediqt
and
claude
committed
Periodic minimum-image neighbor distance for the grid-direct path (ADR-0004)

kunalpuri-prediqt
and
claude
committed
BUILD.md: add MPI + Zoltan + PyZoltan section for distributed parallel runs

kunalpuri-prediqt
and
claude
committed
setup.py: skip the MPI/Zoltan parallel extension when PyZoltan is absent

kunalpuri-prediqt
and
claude
committed
Record production GPU result on RTX PRO 6000 Blackwell (sm_120)

kunalpuri-prediqt
and
claude
committed
Add gpu_perf_sweep.py: cross-GPU throughput sweep for the grid-direct path

kunalpuri-prediqt
and
claude
committed
Add cross-GPU throughput sweep artifact; first full sweep = NVIDIA L40S

kunalpuri-prediqt
and
claude
committed
gpu-sweep: add RTX 5090, RTX PRO 6000 Blackwell, B300 full sweeps + comparison

kunalpuri-prediqt
and
claude
committed
gpu-sweep: add single-CPU-core speedup headline to the cross-GPU artifact

kunalpuri-prediqt
and
claude
committed
gpu-sweep: add figures (throughput, per-step, speedup) + plot script

kunalpuri-prediqt
and
claude
committed
Add $-cost-of-compute lens to cross-GPU sweep

kunalpuri-prediqt
and
claude
committed
Add energy + roofline estimate lenses to cross-GPU sweep; close out

kunalpuri-prediqt
and
claude
committed

Commits on Jun 20, 2026

warp rigid body: device-resident RK2 integration (ADR-0006)
kunalpuri-prediqt
committed

Commits on Jun 21, 2026

Commits on Jul 6, 2026