{
  "abstract": "Level-1 Triggers at the LHC classify collisions within a few microseconds on FPGAs, where a network is a fully unrolled circuit whose cost is area rather than time. We present SCRAMJet, a transformer encoder that replaces multi-head attention with a low-rank projection, trains under per-parameter learned fixed-point quantization, and compiles to Verilog with distributed arithmetic optimization. On the ten-class JetClass dataset it reaches $71.4 \\pm 0.6\\%$ accuracy with 0.157M multiply-accumulates, 6.1 accuracy points above the strongest published resource-constrained model at an eighth of the operations, and meets timing on a Virtex UltraScale+ VU13P at 147 ns using 25.7% of the logic. We also show that resource-aware training silently removes attention channels: one whose learned bit-width reaches zero still carries signal during training but is absent from the compiled circuit, so the loss never sees it go. We call these virtual channels, and a floor on the attention datalane that prevents them is worth 3.1 accuracy points at matched budget. Code and trained checkpoints are available at https://github.com/Aaddy-1/hgq-jetformer.",
  "abstract_shown": true,
  "affiliations": [],
  "article_type": null,
  "arxiv": "2611.00086",
  "author_total": 9,
  "authors": [
    {
      "name": "Aadeesh Sharma"
    },
    {
      "name": "Chang Sun"
    },
    {
      "name": "Lauri Laatu"
    },
    {
      "name": "Arianna Cox"
    },
    {
      "name": "Benedikt Maier"
    },
    {
      "name": "Alexander Tapper"
    },
    {
      "name": "Maria Spiropulu"
    },
    {
      "name": "Wayne Luk"
    },
    {
      "name": "Zhiqiang Que"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/47wthfnxe0qnp.bib",
  "cited_by": 0,
  "doi": "10.48550/arxiv.2611.00086",
  "fulltext_url": null,
  "id": "47wthfnxe0qnp",
  "issns": [],
  "landing_url": "https://arxiv.org/abs/2611.00086",
  "language": null,
  "licence": "LicenseRef-arXiv-nonexclusive-1.0",
  "licence_label": "arXiv licence",
  "licence_source": "article",
  "licence_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
  "metadata_licence": "CC0-1.0",
  "native_subjects": [
    [
      "arxiv",
      "cs.LG"
    ],
    [
      "arxiv",
      "hep-ex"
    ],
    [
      "arxiv",
      "physics.ins-det"
    ]
  ],
  "openalex": null,
  "pdf_url": "https://arxiv.org/pdf/2611.00086",
  "pmcid": null,
  "pmid": null,
  "published": "2026-11-01",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/47wthfnxe0qnp.ris",
  "sources": [
    {
      "first_seen": "2026-11-03T12:20:59Z",
      "native_id": "2611.00086",
      "source": "arxiv",
      "url": "https://arxiv.org/abs/2611.00086"
    }
  ],
  "subjects": [
    {
      "field": "physics",
      "origin": "arxiv:physics.ins-det",
      "subfield": "applied-instrumentation"
    },
    {
      "field": "physics",
      "origin": "arxiv:hep-ex",
      "subfield": "high-energy"
    }
  ],
  "title": "SCRAMJet: An Ultra Low Latency JetClass-Scale Transformer for the HL-LHC Level-1 Trigger",
  "title_text": "SCRAMJet: An Ultra Low Latency JetClass-Scale Transformer for the HL-LHC Level-1 Trigger",
  "type": "preprint",
  "updated_at": "2026-11-05T15:20:18Z",
  "url": "https://research.draconitic.com/paper/47wthfnxe0qnp",
  "venue": "arXiv",
  "versions": []
}