rustytorch

rtx-cfd 3D Stage 1 item 5: piso3.cu + three_d::piso_device::Piso3Device — the PISO step device-resident (predictors compiled without FMA contraction = the host's arithmetic, per-side boundary tables, divergence/correct/add_p kernels with fixed-order p…

Performance Benchmarks #790:Commit d9fd849239 pushed by osobh

main

2026-09-17 16:44:14 +00:00

6s

rtx-cfd three_d: the predictors split into piso_predictor.rs (file-size rule), the w predictor's index helpers cleaned; tests unchanged and green

Documentation #789:Commit e0cbb99343 pushed by osobh

main

2026-09-17 16:34:53 +00:00

1m33s

rtx-cfd three_d: the predictors split into piso_predictor.rs (file-size rule), the w predictor's index helpers cleaned; tests unchanged and green

Performance Benchmarks #787:Commit e0cbb99343 pushed by osobh

main

2026-09-17 16:33:35 +00:00

16s

rtx-cfd 3D Stage 1 item 4: three_d::{FlowField3D, piso_host::Piso3Solver} — the 2D embedded predictor/projection transcribed with the z terms appended (u/v/w predictors, six sides incl. periodic z, TVD, the apertured-ready projection on PoissonProblem…

Documentation #786:Commit 616d2a3394 pushed by osobh

main

2026-09-17 16:33:18 +00:00

1m42s

rtx-cfd 3D Stage 1 item 4: three_d::{FlowField3D, piso_host::Piso3Solver} — the 2D embedded predictor/projection transcribed with the z terms appended (u/v/w predictors, six sides incl. periodic z, TVD, the apertured-ready projection on PoissonProblem…

Performance Benchmarks #784:Commit 616d2a3394 pushed by osobh

main

2026-09-17 16:31:40 +00:00

6s

rtx-cfd 3D Stage 1 item 3: cg3.cu + DeviceCg3 — the f64 CG entirely on the device (run_pcg's control flow: true-residual stop and resync, breakdown guard, mean projection, anchor shift; fixed-order block reductions, no atomics; the device V-cycle as p…

Documentation #783:Commit 2f476a38d5 pushed by osobh

main

2026-09-17 16:20:26 +00:00

1m29s

rtx-cfd 3D Stage 1 item 3: cg3.cu + DeviceCg3 — the f64 CG entirely on the device (run_pcg's control flow: true-residual stop and resync, breakdown guard, mean projection, anchor shift; fixed-order block reductions, no atomics; the device V-cycle as p…

Performance Benchmarks #781:Commit 2f476a38d5 pushed by osobh

main

2026-09-17 16:19:01 +00:00

5s

rtx-cfd 3D Stage 1 item 2: mg3_vcycle.cu (seven-point red-black half-sweep, residual, coarsest; top/bot neighbour arrays carry the periodic wrap), LevelExport3/export_hierarchy3/vcycle_f32_reference3, DeviceVcycle3 (the K=1 sequence); gate 2 HELD: devi…

Documentation #780:Commit 05ed96f383 pushed by osobh

main

2026-09-17 16:15:19 +00:00

5s

rtx-cfd three_d: ignore reasons on the probes (clippy)

Documentation #777:Commit c13b07b68d pushed by osobh

main

2026-09-17 16:12:45 +00:00

11s

rtx-cfd 3D Stage 1 item 1: three_d::{Grid3, poisson} — the 2D Poisson stack transcribed to a seven-point operator (sanitised coefficients, 2×2×2 Galerkin aggregation, (i+j+k)%2 colouring, periodic z, run_pcg line for line, PcgCache3); gate 1 HELD: nz=…

Documentation #774:Commit 0794e1b6db pushed by osobh

main

2026-09-17 16:12:03 +00:00

5s

rtx-cfd/rtx-fsi PERF-2 P3-iii K-stream mode: each lane thread's device operator lives on its own CUDA stream (no rendezvous; set_plane_streams, RTX_FSI2O_PLANE_MODE=streams default | batch); stream-mode unit test digit-identical to the solo apply

Documentation #771:Commit 7e135893a2 pushed by osobh

main

2026-09-17 11:43:56 +00:00

3m5s

rtx-cfd/rtx-fsi PERF-2 P3-iii K-stream mode: each lane thread's device operator lives on its own CUDA stream (no rendezvous; set_plane_streams, RTX_FSI2O_PLANE_MODE=streams default | batch); stream-mode unit test digit-identical to the solo apply

Performance Benchmarks #769:Commit 7e135893a2 pushed by osobh

main

2026-09-17 11:41:04 +00:00

18s

rtx-cfd/rtx-fsi PERF-2 P3-iii: the MarchPlane — lane-table V-cycle kernels (ml_*), the batched apply over K lanes' own operators, the plane rendezvous inside the CG's preconditioner (lane thread-local, CG guard), set_plane_lane/plane_counters; the fsi…

Documentation #768:Commit 95ffde9591 pushed by osobh

main

2026-09-16 22:04:00 +00:00

6s

rtx-cfd/rtx-fsi PERF-2 P3-iii: the MarchPlane — lane-table V-cycle kernels (ml_*), the batched apply over K lanes' own operators, the plane rendezvous inside the CG's preconditioner (lane thread-local, CG guard), set_plane_lane/plane_counters; the fsi…

Performance Benchmarks #766:Commit 95ffde9591 pushed by osobh

main

2026-09-16 22:04:05 +00:00

12s

rtx-cfd legacy GPU tests: the two point-source Poisson tests expect the sign of ∇²φ = f

Documentation #765:Commit 49ccebc863 pushed by osobh

main

2026-09-16 12:37:04 +00:00

49s

rtx-cfd legacy GPU tests: the two point-source Poisson tests expect the sign of ∇²φ = f

Performance Benchmarks #763:Commit 49ccebc863 pushed by osobh

main

2026-09-16 12:36:17 +00:00

6s

rtx-cfd CUDA Poisson kernels: solve ∇²φ = f as documented (the updates subtracted the source with the wrong sign); the Jacobi test's stop is absolute on a source of size 2π²

Documentation #762:Commit eba2f61ddf pushed by osobh

main

2026-09-16 12:34:41 +00:00

6s

rtx-cfd legacy GPU tests: the Jacobi Poisson test stops at 1e-4 (the accuracy check is the pin); rustfmt

Documentation #759:Commit 52fd40f8f2 pushed by osobh

main

2026-09-16 12:32:58 +00:00

23s

rtx-cfd legacy GPU: per-block buffer mutability; the D3Q19 getters extract before borrowing

Documentation #756:Commit 1f77ec637a pushed by osobh

main

2026-09-16 12:31:00 +00:00

6s

rtx-cfd legacy GPU: per-block buffer mutability; the D3Q19 getters extract before borrowing

Performance Benchmarks #754:Commit 1f77ec637a pushed by osobh

main

2026-09-16 12:30:54 +00:00

7s

rtx-cfd legacy GPU: buffers re-borrowed after nested device calls; writing and delegating methods take &mut self

Documentation #753:Commit 8999548aca pushed by osobh

main

2026-09-16 12:30:19 +00:00

6s

rtx-cfd legacy GPU: buffers re-borrowed after nested device calls; writing and delegating methods take &mut self

Performance Benchmarks #751:Commit 8999548aca pushed by osobh

main

2026-09-16 12:30:13 +00:00

7s

rtx-cfd legacy GPU: the kernel-writing methods take &mut self; read-only ones keep a shared borrow

Documentation #750:Commit fd5972c949 pushed by osobh

main

2026-09-16 12:29:38 +00:00

24s

rtx-cfd legacy GPU: the kernel-writing methods take &mut self; read-only ones keep a shared borrow

Performance Benchmarks #748:Commit fd5972c949 pushed by osobh

main

2026-09-16 12:29:14 +00:00

8s

rtx-cfd legacy GPU: kernel outputs were device-buffer CLONES (every write lost) — the buffers are borrowed mutably now; the Poisson kernels keep boundary Dirichlet values; tests: the advection pulse marched 50 steps, the Jacobi budget 20k, the reducti…

Documentation #747:Commit 0de750bad1 pushed by osobh

main

2026-09-16 12:28:53 +00:00

5s

rtx-cfd CUDA kernels: no host includes at all in the NVRTC sources; the float limits defined inline

Documentation #744:Commit f3f6ac21fb pushed by osobh

main

2026-09-16 12:25:16 +00:00

14s

rtx-cfd CUDA kernels: no host includes at all in the NVRTC sources; the float limits defined inline

Performance Benchmarks #742:Commit f3f6ac21fb pushed by osobh

main

2026-09-16 12:25:05 +00:00

7s

rtx-cfd CUDA kernels: no host headers in the NVRTC sources (cuda_runtime.h was never found: every legacy GPU test failed at manager creation)

Documentation #741:Commit 64faa095a2 pushed by osobh

main

2026-09-16 12:23:38 +00:00

27s

rtx-cfd CUDA kernels: no host headers in the NVRTC sources (cuda_runtime.h was never found: every legacy GPU test failed at manager creation)

Performance Benchmarks #739:Commit 64faa095a2 pushed by osobh

main

2026-09-16 12:23:12 +00:00

6s

rtx-cfd legacy GPU modules: the launches that already sat in unsafe blocks are not wrapped twice

Documentation #738:Commit f429bdd415 pushed by osobh

main

2026-09-16 12:21:10 +00:00

6s