{
  "annotation_id": "931278f1-b439-405d-bcfa-3e374b2f38ee",
  "date_created": "2026-03-02T18:00:46.811000Z",
  "date_modified": "2026-03-02T18:00:46.811000Z",
  "file_hash": "5a7ab5ed16122913503bf80ff2d71715bb767b2f5d86f0eb0c51a46fb8b6c963",
  "private": false,
  "record": {
    "abstract": "We describe the offline computing system of the Belle experiment, consisting\nof a computing farm with one thousand IA-32 CPUs. Up to now, the Belle\nexperiment has accumulated more than 120 fb$^{-1}$ of data, which is the world\nlargest $B\\bar{B}$ sample at the $\\Upsilon(4S)$ energy. The data have to be\nprocessed with a single version of reconstruction software and calibration\nconstants to perform precise measurements of $B$ meson decays. In addition,\nMonte Carlo samples three times larger than the real beam data are generated.\nTo fullfill our computing needs, we have constructed the computing system with\n90(300) quad(dual) CPU PC servers from multiple vendors as a central processing\nsystem. The details of this computing system and performance of data processing\nwith the current model are presented.",
    "arxiv_id": "physics/0306120",
    "authors": [
      "I. Adachi",
      "R. Itoh",
      "N. Katayama",
      "T. Tsukamoto",
      "T. Hibino",
      "M. Yokoyama",
      "L. Hinz",
      "F. Ronga"
    ],
    "categories": [
      "physics.ins-det"
    ],
    "title": "Computing System for the Belle Experiment",
    "url": "https://arxiv.org/abs/physics/0306120"
  },
  "schema_id": "dorsal/arxiv",
  "source": {
    "execution_id": "0c995f77-ec95-412b-aeca-d042966678d8",
    "id": "arXiv Dataset IDs",
    "type": "Model",
    "variant": "snapshot-2026-03-01",
    "version": "0.1.0"
  },
  "user_id": 1000002
}