Knowledge Graph Retrieval on PKG Hop 5
101.05QT (ms)LogosKG (Torch-GPU)
Evaluation results
| Method | Links | ||
|---|---|---|---|
| LogosKG (Torch-GPU)Backend=Torch, Hardware=GPU, Scale Variant=Standard2026.04 | 101.05 | 0 | |
| LogosKG (Numba)Backend=Numba, Hardware=CPU, Scale Variant=Standard2026.04 | 204.25 | 0 | |
| GraphBLASHardware=CPU, Scale Variant=Standard2026.04 | 415.43 | 0 | |
| SNAPHardware=CPU, Scale Variant=Standard2026.04 | 446.15 | 0 | |
| igraphHardware=CPU, Scale Variant=Standard2026.04 | 580.91 | 0 | |
| LogosKG (SciPy)Backend=SciPy, Hardware=CPU, Scale Variant=Standard2026.04 | 677.23 | 0 | |
| PyGHardware=GPU, Scale Variant=Standard2026.04 | 735.34 | 0 | |
| DGLHardware=GPU, Scale Variant=Standard2026.04 | 1,099.26 | 0 | |
| NetworkXHardware=CPU, Scale Variant=Standard2026.04 | 1,511.28 | 0 | |
| cuGraphHardware=GPU, Scale Variant=Standard2026.04 | 1,616.55 | 0 | |
| LogosKG (Torch-CPU)Backend=Torch, Hardware=CPU, Scale Variant=Standard2026.04 | 2,207.25 | 0 | |
| graph-toolHardware=CPU, Scale Variant=Standard2026.04 | 2,306.34 | 0.67 | |
| LogosKG (Large-Numba)Backend=Numba, Hardware=CPU, Scale Variant=Large2026.04 | 4,085.72 | 4 | |
| LogosKG (Large-Torch-GPU)Backend=Torch, Hardware=GPU, Scale Variant=Large2026.04 | 4,482.14 | 6 | |
| LogosKG (Large-SciPy)Backend=SciPy, Hardware=CPU, Scale Variant=Large2026.04 | 4,532.75 | 5.33 | |
| LogosKG (Large-Torch-CPU)Backend=Torch, Hardware=CPU, Scale Variant=Large2026.04 | 7,467.56 | 48.67 | |
| Neo4jHardware=CPU, Scale Variant=Standard2026.04 | 10,000 | 100 |