SPB Git

spb/localvm-research Public License

Running LLMs larger than memory on a consumer Mac — falsification-driven research: margin-gated deferred refinement, out-of-core verification on Apple Silicon. TR-01 published.

Python 63.2% JavaScript 23.5% CSS 11.8% Shell 0.9% Makefile 0.5%
2.8 KB · 111 lines json
Raw Blame History
1{2  "experiment": "candidate_01_deferred_refinement",3  "author": "Simon-Pierre Boucher",4  "contact": "contact@spboucher.ai",5  "manifest": {6    "author": "Simon-Pierre Boucher",7    "contact": "contact@spboucher.ai",8    "project": "localvm-research",9    "collected_utc": "2026-08-12T05:09:16.535264+00:00",10    "chip": {11      "brand": "Apple M5 Max",12      "arch": "arm64",13      "cores_total": 18,14      "cores_performance": 6,15      "cores_efficiency": 12,16      "gpu_cores": 4017    },18    "memory": {19      "unified_bytes": 51539607552,20      "unified_gb": 48.0,21      "pagesize": 1638422    },23    "ssd": {24      "model": "APPLE SSD AP2048Z",25      "size": "2 TB",26      "smart_status": "Verified"27    },28    "os": {29      "product": "macOS",30      "version": "27.0",31      "build": "26A5388g",32      "kernel": "27.0.0"33    },34    "software": {35      "python": "3.14.4",36      "mlx": "0.32.0",37      "mlx_lm": "0.31.3",38      "torch": null,39      "numpy": "2.5.2"40    },41    "git": {42      "commit": "74574f3169ebc78c217e870cbb54b9696bfa0ac7",43      "dirty_tree": true44    },45    "thermal_level_at_collect": null46  },47  "config": {48    "build": false,49    "per_domain": 1,50    "max_tokens": 48,51    "window": 32,52    "taus": "2.0"53  },54  "model": "mlx-community/Qwen3-1.7B-bf16",55  "q4_resident_bytes": 968080210,56  "q8_stream_bytes": 1828306411,57  "baselines": {58    "pure_q4": {59      "tokens_per_s": 62.069495034236134,60      "fidelity_vs_q8_mean": 0.760416666666666661    },62    "pure_q8": {63      "tokens_per_s": 48.93987602867763,64      "fidelity_vs_q8_mean": 1.065    }66  },67  "runs": [68    {69      "mode": "margin",70      "tau": 2.0,71      "window": 32,72      "fidelity_vs_q8_mean": 0.7204084820801689,73      "tokens_per_s": 31.77510653276482,74      "deferral_rate": 0.2922374429223744,75      "rollback_rate": 0.045662100456621,76      "sweeps_per_100tok": 6.392694063926941,77      "sweep_latency_s_mean": 0.04829824392839198,78      "logical_verify_bytes_per_token": 116878035.4063927,79      "raw": {80        "tokens": 219,81        "deferred": 64,82        "sweeps": 14,83        "rollbacks": 10,84        "sweep_s": 0.6761754149974877,85        "gen_s": 6.216012335000414,86        "logical_bytes": 2559628975487      }88    },89    {90      "mode": "verify-all",91      "tau": 2.0,92      "window": 32,93      "fidelity_vs_q8_mean": 0.7684465642771071,94      "tokens_per_s": 27.02619478039896,95      "deferral_rate": 0.3604060913705584,96      "rollback_rate": 0.06598984771573604,97      "sweeps_per_100tok": 7.106598984771574,98      "sweep_latency_s_mean": 0.04714166964289949,99      "logical_verify_bytes_per_token": 129930404.8426396,100      "raw": {101        "tokens": 197,102        "deferred": 71,103        "sweeps": 14,104        "rollbacks": 13,105        "sweep_s": 0.6599833750005928,106        "gen_s": 6.629241083001034,107        "logical_bytes": 25596289754108      }109    }110  ]111}