{
  "schemaVersion": 1,
  "study": "semafold",
  "reviewed": "2026-09-25",
  "evidenceType": "extracted-source-report",
  "reproducedInThisWebsiteReview": false,
  "repository": "https://github.com/mindtro/semafold",
  "commit": "096e6dc497bcc87a1ead2466c15841f063df1519",
  "libraryVersion": "0.2.0",
  "benchmarkVersion": "0.1.0",
  "benchmarkDate": "2026-03-31",
  "scope": "Deterministic synthetic vector and KV tensor storage benchmarks. Not application latency, retrieval recall or real-model quality measurements. Compression and KV codec APIs remain preview.",
  "sources": [
    {
      "name": "benchmarks/turboquant_benchmark_report.md",
      "sha256": "ef384c74195d3f0944b36b1d2cf75baa78ca77e6e29009faa1f08ed3d648f1ad"
    },
    {
      "name": "STABILITY.md",
      "sha256": "f085264ec4f6c0a0b3e962e8cccea470fd6e11fdb501f3492b4b4865ae6885e7"
    },
    {
      "name": "CHANGELOG.md",
      "sha256": "b52cccda6dc65f2d5f8de5f31d7c44a634d1250dcd61c5b1cd1c5ebd46b603f6"
    },
    {
      "name": "pyproject.toml",
      "sha256": "b5ed5776e2bdde03a2ff6b62176ed687993b9a7bc39e4e74721f49c8763950ce"
    }
  ],
  "tables": [
    {
      "title": "Reconstruction-Oriented Storage (`TurboQuantMSE`)",
      "source": "benchmarks/turboquant_benchmark_report.md",
      "rows": [
        [
          "Setting",
          "Dense float32",
          "Dense fp16/bf16",
          "Semafold Artifact",
          "Saved vs float32",
          "Smaller vs float32",
          "Smaller vs fp16",
          "Ratio vs float32",
          "Ratio vs fp16",
          "Quality Proxy"
        ],
        [
          "2-bit MSE",
          "786,432 B",
          "393,216 B",
          "50,077 B",
          "736,355 B",
          "93.63%",
          "87.26%",
          "15.70x",
          "7.85x",
          "observed MSE 0.1178"
        ],
        [
          "3-bit MSE",
          "786,432 B",
          "393,216 B",
          "74,738 B",
          "711,694 B",
          "90.50%",
          "80.99%",
          "10.52x",
          "5.26x",
          "observed MSE 0.0348"
        ],
        [
          "4-bit MSE",
          "786,432 B",
          "393,216 B",
          "99,484 B",
          "686,948 B",
          "87.35%",
          "74.70%",
          "7.91x",
          "3.95x",
          "observed MSE 0.0098"
        ]
      ]
    },
    {
      "title": "Inner-Product-Oriented Storage (`TurboQuantProd`)",
      "source": "benchmarks/turboquant_benchmark_report.md",
      "rows": [
        [
          "Setting",
          "Dense float32",
          "Dense fp16/bf16",
          "Semafold Artifact",
          "Saved vs float32",
          "Smaller vs float32",
          "Smaller vs fp16",
          "Ratio vs float32",
          "Ratio vs fp16",
          "Quality Proxy"
        ],
        [
          "2 total bits",
          "786,432 B",
          "393,216 B",
          "50,638 B",
          "735,794 B",
          "93.56%",
          "87.12%",
          "15.53x",
          "7.77x",
          "mean abs dot error 0.0156"
        ],
        [
          "3 total bits",
          "786,432 B",
          "393,216 B",
          "75,255 B",
          "711,177 B",
          "90.43%",
          "80.86%",
          "10.45x",
          "5.23x",
          "mean abs dot error 0.0089"
        ],
        [
          "4 total bits",
          "786,432 B",
          "393,216 B",
          "99,916 B",
          "686,516 B",
          "87.30%",
          "74.59%",
          "7.87x",
          "3.94x",
          "mean abs dot error 0.0048"
        ]
      ]
    },
    {
      "title": "Representative Cache-Shaped Tensor Result",
      "source": "benchmarks/turboquant_benchmark_report.md",
      "rows": [
        [
          "Workload",
          "Shape",
          "Dense float32",
          "Dense fp16/bf16",
          "Semafold Combined Artifact",
          "Smaller vs float32",
          "Smaller vs fp16",
          "Ratio vs float32",
          "Ratio vs fp16/bf16"
        ],
        [
          "KV block compression",
          "(4, 8, 256, 128)",
          "8,388,608 B",
          "4,194,304 B",
          "885,734 B",
          "89.44%",
          "78.88%",
          "9.47x",
          "4.74x"
        ]
      ]
    },
    {
      "title": "K / V Split",
      "source": "benchmarks/turboquant_benchmark_report.md",
      "rows": [
        [
          "Component",
          "Artifact Bytes"
        ],
        [
          "Keys (TurboQuantProd)",
          "459,255 B"
        ],
        [
          "Values (TurboQuantMSE)",
          "426,479 B"
        ],
        [
          "Combined",
          "885,734 B"
        ]
      ]
    },
    {
      "title": "Small-Block Caveat",
      "source": "benchmarks/turboquant_benchmark_report.md",
      "rows": [
        [
          "Workload",
          "Shape",
          "Dense float32",
          "Dense fp16",
          "Semafold Combined Artifact",
          "Smaller vs float32",
          "Smaller vs fp16",
          "Ratio vs float32",
          "Ratio vs fp16"
        ],
        [
          "Tiny KV block",
          "(2, 2, 6, 16)",
          "3,072 B",
          "1,536 B",
          "1,550 B",
          "49.54%",
          "-0.91%",
          "1.98x",
          "0.99x"
        ]
      ]
    }
  ]
}