engineering/lumbda
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

bend catalog Wave 3 — 15 surveyed forms (2026-06-05)

Browse source 
Adds 15 new GPU primitive forms across domains not covered by
Waves 1-2: ZK/SNARK provers, pairing crypto, tensor network
contraction, sparse LA, CV primitives, numerical solvers,
generic belief propagation, MD/CFD, convex opt, DSP beyond cuFFT,
DB aggregations, graph theory beyond triangle/PageRank.

Top hits by relevance to our missions:

  * cuda-mfcc-spectral (97x) — punters-cc, BT-DISC forensics, real-time CC pipeline. unsandbox HIGH.
  * cuda-batched-lp-simplex (95x over CPLEX) — Prime Mission workstation-to-queue balancing. unsandbox HIGH.
  * cuda-betweenness-centrality-weighted (30-150x) — workaholic-node detection on dependency DAG, MOAD-0001 model. undefect HIGH.
  * cuda-hungarian-lap (400M-var LAP ~13s) — workstation balancing. unsandbox HIGH.
  * cuda-msm-bls12-381 (27.86x) — Pippenger shares branchless modmul shape with our reversible secp256k1 inner loop. ECDSA HIGH.
  * cuda-tensornet-contract (tensor QR 100x vs Xeon 8480+) — alt to stabilizer/kickmix sim path; MPS/PEPS evaluates reversible secp256k1 circuits beyond Clifford. ECDSA HIGH.
  * cuda-loopy-bp-mrf (45x) — LBP substrate for FuzzingBrain probabilistic analysis. undefect HIGH.
  * cuda-ega-gpu-aggregation (6-29x) — defect-corpus aggregation at planetary scale; telemetry queue aggregation. undefect HIGH + unsandbox HIGH.

Plus highlight numbers: FluidX3D LBM 100-200x vs ANSYS Fluent,
CUDASW++4.0 5.71 TCUPS on H100 for sequence alignment, ICICLE-Snark
the fastest Groth16 today.

Wave 3 filter-outs noted: AMGX, GROMACS, NVOFA, junction-tree BP,
batched L-BFGS — all below 10x or insufficiently general.

CATALOG.md mirrors the table for in-repo reading; bend.html
renders the public version with reference URLs.
This commit is contained in:
russell@unturf.com 2026-06-05 13:18:06 -04:00
parent a17c114263
commit a33ae4331d
No known key found for this signature in database
2 changed files with 65 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

33
examples/cuda-fanout/CATALOG.md
View file

@ -141,6 +141,39 @@ run on our fleet yet.
- cuda-batched-mcts — 25-40x on Go-style rollouts; wrong shape for our search
- cuda-faiss-ann — 5-12x; no embedding workload today
## Wave 3 — surveyed 2026-06-05
15 additional forms covering ZK/SNARK provers, pairing crypto,
tensor network contraction, sparse LA, CV primitives, numerical
solvers, generic BP, MD/CFD, convex opt, DSP beyond cuFFT, DB
aggregations, graph theory beyond triangle/PageRank.
| form | speedup / throughput | hardware | relevance |
|-----------------------------------|------------------------------------------------|---------------------|-----------|
| cuda-fluidx3d-lbm | 100-200x vs ANSYS Fluent; 8799 MLUPS A100 | A100 | unsandbox MEDIUM (HPC reproducibility) |
| cuda-mfcc-spectral | ~97x CPU MFCC; STFT ~75x via cuSignal | GTX 580+ | unsandbox HIGH (punters-cc, BT-DISC forensics) |
| cuda-batched-lp-simplex | 95x over CPLEX; 5x over GLPK on 100K LPs | GTX 980-class | unsandbox HIGH (resource scheduling, Prime Mission) |
| cuda-betweenness-centrality-weighted | 30-150x warp-centric weighted BC | GTX+ | undefect HIGH (workaholic-node detection, MOAD-0001) |
| cuda-cudasift-orb-ransac | ~60x SIFT; 1.2ms on GTX 1060; ORB 11.3x | GTX 1060+ | unsandbox MEDIUM (visual evidence) |
| cuda-cudasw-gasal2 | CUDASW++4.0 16.2x; 5.71 TCUPS H100; GASAL2 750x | H100 | undefect MEDIUM (binary-diff patch-similarity) |
| cuda-loopy-bp-mrf | 45x over CPU LBP for stereo MRF | GTX 280+ | undefect HIGH (FuzzingBrain probabilistic analysis) |
| cuda-sgm-stereo | 42 fps @ 640x480, 128 disparities Tegra X1 | Tegra X1+ | unsandbox MEDIUM (edge node profile) |
| cuda-hungarian-lap | 400M-var LAP in ~13s; 10-50x class | NVIDIA | unsandbox HIGH (workstation-to-queue balancing); undefect MEDIUM |
| cuda-msm-bls12-381 | 27.86x over Pippenger RELIC AVX; 60% of Groth16 | A100 / 4090 | ECDSA HIGH (Pippenger shares branchless modmul shape) |
| cuda-pdwt-lifting | 15.9x over best CPU DWT | GTX / Tesla | unsandbox MEDIUM (audio-IPC analysis) |
| cuda-tensornet-contract | 8-20x vs CuPy; tensor QR ~100x vs Xeon 8480+ | A100 | ECDSA HIGH (alt to stabilizer/kickmix) |
| cuda-ega-gpu-aggregation | 6.45-29.12x multi-pass; group-by hash 19.4x | NVIDIA | undefect HIGH (corpus aggregation); unsandbox HIGH (telemetry) |
| cuda-bicgstab-ilu-spmv | SpTRSV 10.7x; ILU0 BiCGSTAB 3.2x; GMRES 6.9x | V100 / MI210 | ECDSA MEDIUM (sparse LA over GF(p)); undefect MEDIUM |
| cuda-icicle-snark-groth16 | Fastest Groth16 today; Mina 3x libsnark | RTX 4090 / A100 | ECDSA MEDIUM; undefect MEDIUM (zk-prover defect scanning) |
### Wave 3 filtered out (below threshold)
- AMGX algebraic multigrid: 2-5x — below 10x cutoff
- GROMACS GPU: 2-3x speedup vs CPU — below threshold
- NVIDIA Optical Flow SDK: 7-10x borderline, dedicated hardware unit
- Junction-tree BP per-message: 0.68-9.18x — below threshold
- Batched L-BFGS: 134x reported but single mesh-size case, not generalized
## Build order — recommended 3-form runway
1. **A: `cuda-secp256k1-batched-mul`** — biggest immediate win. Drop VanitySearch's CUDA secp256k1 kernel as a BSHK worker; lumbda gets a 6.5 Gkeys/s point-mul oracle. Every candidate validation rides this form. Hardware already owned (3090 + 4090).