{
  "annotation_id": "839655d4-95e2-4a31-8e54-2e4c9390b090",
  "date_created": "2026-02-17T05:53:23.785000Z",
  "date_modified": "2026-02-17T05:53:23.785000Z",
  "file_hash": "fa3bdc6b04cd6bbea51b79e77f11c96cc88b7a69a4566b5904c5e48b4f4d54ad",
  "private": false,
  "record": {
    "abstract": "Neuromorphic computers hold the potential to vastly improve the speed and efficiency of a wide range of computational kernels with their asynchronous, compute-memory co-located, spatially distributed, and scalable nature. However, performance models that are simple yet sufficiently expressive to predict runtime on actual neuromorphic hardware are lacking, posing a challenge for researchers and developers who strive to design fast algorithms and kernels. As breaking the memory bandwidth wall of conventional von-Neumann architectures is a primary neuromorphic advantage, modeling communication time is especially important. At the same time, modeling communication time is difficult, as complex congestion patterns arise in a heavily-loaded Network-on-Chip. In this work, we introduce the first max-affine lower-bound runtime model -- a multi-dimensional roofline model -- for Intel\u0027s Loihi 2 neuromorphic chip that quantitatively accounts for both compute and communication based on a suite of microbenchmarks. Despite being a lower-bound model, we observe a tight correspondence (Pearson correlation coefficient greater than or equal to 0.97) between our model\u0027s estimated runtime and the measured runtime on Loihi 2 for a neural network linear layer, i.e., matrix-vector multiplication, and for an example application, a Quadratic Unconstrained Binary Optimization solver. Furthermore, we derive analytical expressions for communication-bottlenecked runtime to study scalability of the linear layer, revealing an area-runtime tradeoff for different spatial workload configurations with linear to superliner runtime scaling in layer size with a variety of constant factors. Our max-affine runtime model helps empower the design of high-speed algorithms and kernels for Loihi 2.",
    "arxiv_id": "2601.10035",
    "authors": [
      "Jonathan Timcheck",
      "Alessandro Pierro",
      "Sumit Bam Shrestha"
    ],
    "categories": [
      "cs.NE"
    ],
    "license": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
    "title": "A Compute and Communication Runtime Model for Loihi 2",
    "url": "https://arxiv.org/abs/2601.10035",
    "version": "v1"
  },
  "schema_id": "dorsal/arxiv",
  "source": {
    "execution_id": "4461f2e9-a079-47b3-ad20-2c9b43972aee",
    "id": "arXiv Dataset",
    "type": "Model",
    "variant": "snapshot-2026-01-17",
    "version": "0.1.0"
  },
  "user_id": 1000002
}