Paper 08 · May 2026

GTC Runtime: Measured Cache Coverage and Batch Jacobi

William Ken Ohara Stewart

HyperTensor Project · Extended version · TeX source

Abstract

Empirical validation of the Geodesic Trajectory Cache (GTC) proposed in Paper IV. On three model scales (135M–1.5B parameters), GTC achieves cache hit rates of 90.4–91.5% at a 25%-fraction coverage budget — scale-invariant within ±0.5%. Batch Jacobi evaluation reaches 97× speedup at B=10. GTC is 15.5× faster than disk-based RAG. Compressed trajectory records occupy 5.96 KB each. End-to-end speculative decode with GTC drafts achieves 76.5 tok/s (1.53× baseline).

1. Key Results

MetricValue
Cache coverage (3-model avg)91.0% at 25% budget
Batch Jacobi peak97.9× at B=10
vs. disk-based RAG15.5× faster
Record size5.96 KB (rank-5 exact)
Query latency30.9 μs
OTT end-to-end76.5 tok/s (1.53× baseline)

2. Scale Invariance

Cache coverage is independent of model size: 91.0% (135M), 90.4% (360M), 91.5% (1.5B) — within ±0.5% across a 33× parameter range. The manifold structure is scale-invariant, confirming the intrinsic dimension hypothesis from Paper IV.

References

  1. Stewart, W.K.O. Organic Training Theory. HyperTensor Paper IV, 2026.
  2. Stewart, W.K.O. GRC Attention Compression. HyperTensor Paper I, 2026.