{
  "schemaVersion": 1,
  "id": "atlas-lesson-crg-communication-graph",
  "title": "Communication is the realization medium",
  "category": "learning",
  "format": "Atlas lesson",
  "description": "Maps computational state, data movement and control into a graph of physical interfaces.",
  "canonicalUrl": "https://semiaifoundry.com/learn/atlas/crg-communication-graph/",
  "publishedDate": null,
  "revisionDate": "2026-10-07",
  "author": null,
  "topic": null,
  "series": null,
  "pursuitIds": [
    "computation-and-realization"
  ],
  "license": null,
  "rightsUrl": "https://semiaifoundry.com/ai-atlas/licenses.html",
  "version": "1.3.0",
  "status": "Published educational lesson",
  "evidenceScope": "The edge equation is a bounded teaching model. Shannon’s source addresses communication limits; it does not by itself specify a chip architecture.",
  "contentCoverage": "published-lesson",
  "artifactUrl": "https://semiaifoundry.com/ai-atlas/lesson-data/crg-communication-graph-5971d22cd051d5a1.json",
  "interactiveUrl": "https://semiaifoundry.com/ai-atlas/#crg-communication-graph",
  "related": [],
  "supersedes": [],
  "supersededBy": [],
  "contentHash": "929677ecb3030bdc8e0e1c5889e197ba86dfba746f7e836b55ec5147e494d921",
  "contentLength": 5416,
  "jsonUrl": "https://semiaifoundry.com/agents/publications/atlas-lesson-crg-communication-graph.json",
  "markdownUrl": "https://semiaifoundry.com/agents/publications/atlas-lesson-crg-communication-graph.md",
  "sections": [
    {
      "id": "section-1",
      "title": "Communication is the realization medium",
      "start": 16,
      "end": 383
    },
    {
      "id": "section-2",
      "title": "The mechanism",
      "start": 383,
      "end": 1472
    },
    {
      "id": "section-3",
      "title": "Science and assumptions",
      "start": 1472,
      "end": 1500
    },
    {
      "id": "section-4",
      "title": "A communication edge",
      "start": 1500,
      "end": 2609
    },
    {
      "id": "section-5",
      "title": "Engineering example",
      "start": 2609,
      "end": 2823
    },
    {
      "id": "section-6",
      "title": "Distinctions and tradeoffs",
      "start": 2823,
      "end": 3261
    },
    {
      "id": "section-7",
      "title": "Specifications and interfaces",
      "start": 3261,
      "end": 3531
    },
    {
      "id": "section-8",
      "title": "Interfaces",
      "start": 3531,
      "end": 3678
    },
    {
      "id": "section-9",
      "title": "Check your understanding",
      "start": 3678,
      "end": 3707
    },
    {
      "id": "section-10",
      "title": "Why does the CRG include communication beside computation?",
      "start": 3707,
      "end": 4141
    },
    {
      "id": "section-11",
      "title": "Evidence and scope",
      "start": 4141,
      "end": 4483
    },
    {
      "id": "section-12",
      "title": "Sources",
      "start": 4483,
      "end": 4999
    },
    {
      "id": "section-13",
      "title": "Continue exploring",
      "start": 4999,
      "end": 5022
    },
    {
      "id": "section-14",
      "title": "Foundations",
      "start": 5022,
      "end": 5233
    },
    {
      "id": "section-15",
      "title": "Related lessons",
      "start": 5233,
      "end": 5416
    }
  ],
  "citations": [
    {
      "title": "Explore in Atlas",
      "url": "https://semiaifoundry.com/ai-atlas/#crg-communication-graph"
    },
    {
      "title": "Semi AI Foundry — Communication Theory of Computation Realization",
      "url": "https://semiaifoundry.com/updates/communication-theory-of-computation-realization/"
    },
    {
      "title": "Claude Shannon — A Mathematical Theory of Communication",
      "url": "https://people.math.harvard.edu/~ctm/home/text/others/shannon/entropy/entropy.pdf"
    },
    {
      "title": "Published lesson data",
      "url": "https://semiaifoundry.com/ai-atlas/lesson-data/crg-communication-graph-5971d22cd051d5a1.json"
    },
    {
      "title": "Atlas releases and licensing",
      "url": "https://semiaifoundry.com/ai-atlas/licenses.html"
    },
    {
      "title": "The computation-to-realization gap",
      "url": "https://semiaifoundry.com/learn/atlas/crg-problem-statement/"
    },
    {
      "title": "On-chip interconnect fabric",
      "url": "https://semiaifoundry.com/learn/atlas/on-chip-network/"
    },
    {
      "title": "NVLink GPU interface",
      "url": "https://semiaifoundry.com/learn/atlas/nvlink-port/"
    },
    {
      "title": "Collective communication library",
      "url": "https://semiaifoundry.com/learn/atlas/nccl/"
    }
  ],
  "recordHash": "1a76d3056441287734d3eef11992543ee1ed0ba1cedf7aee011de4ea0b7261f2",
  "contentFormat": "text/markdown",
  "content": "· Atlas v1.3.0\n\n# Communication is the realization medium\n\nMaps computational state, data movement and control into a graph of physical interfaces.\n\n[Explore in Atlas](https://semiaifoundry.com/ai-atlas/#crg-communication-graph)Signal, wire, link, fabric and workload graph\n\nLearning objectiveDraw a communication edge with a payload, receiver, timing boundary and error condition.\n\n## The mechanism\n\nEvery physical implementation communicates: gates exchange signals, memories exchange addresses and data, dies exchange packets, and systems exchange work and checkpoints. The computation-realization graph treats these transfers as first-class objects with source, destination, encoding, rate, latency, energy and error conditions. This makes data movement visible beside arithmetic. It also explains why communication theory is useful to chip design: a representation is only operational when a receiver can distinguish it within the available noise, timing and power margins. The graph can be refined from a conceptual edge to a routed wire or protocol transaction without changing its teaching role.\n\n1.  Represent a computational dependency as a producer-to-consumer edge.\n2.  Assign an encoding and distinguishability condition to the edge.\n3.  Add rate, latency, energy and error budgets at the chosen boundary.\n4.  Map logical edges to physical paths, buffers, interfaces and protocols.\n5.  Measure delivered useful information rather than nominal link capacity.\n\n## Science and assumptions\n\n### A communication edge\n\n$$\nt_{\\mathrm{edge}}=\\alpha+\\frac{M}{\\beta},\\qquad E_{\\mathrm{edge}}=M e_{\\mathrm{bit}}\n$$\n\nEdge time is startup latency plus transferred bits divided by effective bit bandwidth. Edge energy is transferred bits times effective joules per bit.\n\nM is bits in both relations. If a payload is given in bytes, convert it before combining it with bit bandwidth or joules per bit; the energy term omits any separately unaccounted startup energy.\n\nSymbols and units\n\nt\\_edge\n\nModeled edge communication time\n\ns\n\nα\n\nFixed edge startup latency\n\ns\n\nM\n\nBits transferred at the defined edge boundary\n\nbits\n\nβ\n\nEffective bit-transfer bandwidth at that boundary\n\nbits/s\n\nE\\_edge\n\nModeled energy associated with those transferred bits\n\nJ\n\ne\\_bit\n\nEffective energy per transferred bit under the stated convention\n\nJ/bit\n\nA simple edge model separates startup latency α from payload transfer M/β and scales energy with transferred bits and an effective energy-per-bit term.\n\nAssumptions Homogeneous effective bandwidth and energy; protocol overhead, contention and topology require additional terms.\n\n## Engineering example\n\nFor a 1 GB collective payload over a 100 GB/s effective path with 5 μs startup, the payload term is about 10 ms; a topology or contention change can dominate the same arithmetic operation.\n\n## Distinctions and tradeoffs\n\nA link labeled high bandwidth does not guarantee useful workload throughput when latency, contention, serialization or receiver capacity binds.\n\nMore communication can improve locality or redundancy while increasing wires, buffers, power and verification surface.\n\nAttention, all-reduce, KV-cache traffic and checkpointing are communication graphs whose edges can bind an AI system before arithmetic does.\n\n## Specifications and interfaces\n\nEdge fields\n\nPayload, encoding, rate, latency, energy, error\n\nValues require a declared boundary and time period.\n\nNode fields\n\nState, storage, transform or control\n\nA node may be logical or physical; the model labels the abstraction.\n\n### Interfaces\n\nState producer Input\n\nEncoded data, control or measurement\n\nState receiver Output\n\nAccepted data with timing and error conditions\n\n## Check your understanding\n\n### Why does the CRG include communication beside computation?\n\n1.  Because arithmetic never consumes energy\n2.  Because physical states must be transferred and distinguished between operations\n3.  Because all chips use the same network topology\n\nAnswer and reasoning\n\nBecause physical states must be transferred and distinguished between operations\n\nThe receiver and the path are part of a physical implementation of a computation.\n\n## Evidence and scope\n\nThe edge equation is a bounded teaching model. Shannon’s source addresses communication limits; it does not by itself specify a chip architecture.\n\nUse the manufacturing investigation to trace compatible inputs, assembly and operating constraints; no current standalone calculator models this lesson’s full mechanism.\n\n## Sources\n\n-   [Semi AI Foundry — Communication Theory of Computation Realization](https://semiaifoundry.com/updates/communication-theory-of-computation-realization/)\n-   [Claude Shannon — A Mathematical Theory of Communication](https://people.math.harvard.edu/~ctm/home/text/others/shannon/entropy/entropy.pdf)\n\n[Published lesson data](https://semiaifoundry.com/ai-atlas/lesson-data/crg-communication-graph-5971d22cd051d5a1.json) · [Atlas releases and licensing](https://semiaifoundry.com/ai-atlas/licenses.html)\n\n## Continue exploring\n\n### Foundations\n\n-   [The computation-to-realization gap](https://semiaifoundry.com/learn/atlas/crg-problem-statement/)\n-   [On-chip interconnect fabric](https://semiaifoundry.com/learn/atlas/on-chip-network/)\n\n### Related lessons\n\n-   [NVLink GPU interface](https://semiaifoundry.com/learn/atlas/nvlink-port/)\n-   [Collective communication library](https://semiaifoundry.com/learn/atlas/nccl/)"
}