{
  "schemaVersion": 1,
  "id": "atlas-lesson-crg-package-system-closure",
  "title": "Package and system preserve the contract",
  "category": "learning",
  "format": "Atlas lesson",
  "description": "Carries chip interfaces through bonding, boards, racks, facilities and runtime services.",
  "canonicalUrl": "https://semiaifoundry.com/learn/atlas/crg-package-system-closure/",
  "publishedDate": null,
  "revisionDate": "2026-10-07",
  "author": null,
  "topic": null,
  "series": null,
  "license": null,
  "rightsUrl": "https://semiaifoundry.com/ai-atlas/licenses.html",
  "version": "1.3.0",
  "status": "Published educational lesson",
  "evidenceScope": "Public product and package pages describe representative boundaries; exact fleet configurations, allocation and service telemetry remain configuration-specific.",
  "contentCoverage": "published-lesson",
  "artifactUrl": "https://semiaifoundry.com/ai-atlas/lesson-data/crg-package-system-closure-00378de589c06262.json",
  "interactiveUrl": "https://semiaifoundry.com/ai-atlas/#crg-package-system-closure",
  "related": [],
  "supersedes": [],
  "supersededBy": [],
  "contentHash": "caf8f31a61a4a2aa88c30b0daadfa3c1034b0a08a5e49bbf7896fd9e24f2b71b",
  "contentLength": 5443,
  "jsonUrl": "https://semiaifoundry.com/agents/publications/atlas-lesson-crg-package-system-closure.json",
  "markdownUrl": "https://semiaifoundry.com/agents/publications/atlas-lesson-crg-package-system-closure.md",
  "sections": [
    {
      "id": "section-1",
      "title": "Package and system preserve the contract",
      "start": 16,
      "end": 392
    },
    {
      "id": "section-2",
      "title": "The mechanism",
      "start": 392,
      "end": 1404
    },
    {
      "id": "section-3",
      "title": "Science and assumptions",
      "start": 1404,
      "end": 1432
    },
    {
      "id": "section-4",
      "title": "Useful-work boundary",
      "start": 1432,
      "end": 2540
    },
    {
      "id": "section-5",
      "title": "Engineering example",
      "start": 2540,
      "end": 2776
    },
    {
      "id": "section-6",
      "title": "Distinctions and tradeoffs",
      "start": 2776,
      "end": 3202
    },
    {
      "id": "section-7",
      "title": "Specifications and interfaces",
      "start": 3202,
      "end": 3482
    },
    {
      "id": "section-8",
      "title": "Interfaces",
      "start": 3482,
      "end": 3645
    },
    {
      "id": "section-9",
      "title": "Check your understanding",
      "start": 3645,
      "end": 3674
    },
    {
      "id": "section-10",
      "title": "What does system closure add after tapeout?",
      "start": 3674,
      "end": 4061
    },
    {
      "id": "section-11",
      "title": "Evidence and scope",
      "start": 4061,
      "end": 4417
    },
    {
      "id": "section-12",
      "title": "Sources",
      "start": 4417,
      "end": 4837
    },
    {
      "id": "section-13",
      "title": "Continue exploring",
      "start": 4837,
      "end": 4860
    },
    {
      "id": "section-14",
      "title": "Foundations",
      "start": 4860,
      "end": 5171
    },
    {
      "id": "section-15",
      "title": "Related lessons",
      "start": 5171,
      "end": 5443
    }
  ],
  "citations": [
    {
      "title": "Explore in Atlas",
      "url": "https://semiaifoundry.com/ai-atlas/#crg-package-system-closure"
    },
    {
      "title": "TSMC CoWoS technology",
      "url": "https://3dfabric.tsmc.com/english/dedicatedFoundry/technology/cowos.htm"
    },
    {
      "title": "NVIDIA DGX GB200 hardware guide",
      "url": "https://docs.nvidia.com/dgx/dgxgb200-user-guide/hardware.html"
    },
    {
      "title": "Published lesson data",
      "url": "https://semiaifoundry.com/ai-atlas/lesson-data/crg-package-system-closure-00378de589c06262.json"
    },
    {
      "title": "Atlas releases and licensing",
      "url": "https://semiaifoundry.com/ai-atlas/licenses.html"
    },
    {
      "title": "From architecture contract to a chip",
      "url": "https://semiaifoundry.com/learn/atlas/crg-chip-design-realization/"
    },
    {
      "title": "Package substrate",
      "url": "https://semiaifoundry.com/learn/atlas/package-substrate/"
    },
    {
      "title": "Integrated accelerator rack: GB200 NVL72 example",
      "url": "https://semiaifoundry.com/learn/atlas/rack/"
    },
    {
      "title": "GPU logic dies",
      "url": "https://semiaifoundry.com/learn/atlas/gpu-die/"
    },
    {
      "title": "High-bandwidth memory stack",
      "url": "https://semiaifoundry.com/learn/atlas/hbm-stack/"
    },
    {
      "title": "Distributed checkpoint and recovery state",
      "url": "https://semiaifoundry.com/learn/atlas/checkpoint/"
    }
  ],
  "recordHash": "ebedff7f2bae37adeb09263552dbac972dd8ec6d51396749a6397fc3c8363382",
  "contentFormat": "text/markdown",
  "content": "· Atlas v1.3.0\n\n# Package and system preserve the contract\n\nCarries chip interfaces through bonding, boards, racks, facilities and runtime services.\n\n[Explore in Atlas](https://semiaifoundry.com/ai-atlas/#crg-package-system-closure)Package → node → rack → hall → workload\n\nLearning objectiveTrace a chip interface through package and infrastructure boundaries to a retained workload result.\n\n## The mechanism\n\nThe chip is one participant in a larger communication graph. Bumps, interposers, substrates, power delivery, thermal interfaces, host memory, network fabrics and software runtimes determine whether the die can sustain its intended work. A package can pass electrical test yet fail a thermal or mechanical boundary; a rack can contain tested parts yet lack commissioned power or a compatible data path. CRG therefore continues past tapeout to package and system closure, where interface budgets, workload traces, recovery and useful retained output are measured together.\n\n1.  Match die interfaces to package routing, bonding, power and thermal budgets.\n2.  Test the package and integrate it on a board with host, memory and I/O paths.\n3.  Commission racks, fabric, cooling and facility services at declared boundaries.\n4.  Load data and schedule a workload through compiler, runtime and communication paths.\n5.  Measure retained work, errors, recovery and quality over the operating interval.\n\n## Science and assumptions\n\n### Useful-work boundary\n\n$$\nW_{\\mathrm{useful}}=W_{\\mathrm{executed}}q(1-f_{\\mathrm{lost}})\n$$\n\nUseful work is executed work times its declared acceptance fraction times the fraction retained after modeled losses.\n\nThe two factors require a conditional or otherwise justified separable accounting convention. Do not assume independent quality and interruption outcomes or double-count rejected work. W here denotes work quantity, not watts.\n\nSymbols and units\n\nW\\_useful\n\nAccepted retained work at the declared task boundary\n\ndeclared work units\n\nW\\_executed\n\nExecuted work before the modeled acceptance/retention reductions\n\nsame work units\n\nq\n\nQuality/acceptance fraction under the declared conditional or separable accounting\n\nfraction\n\nf\\_lost\n\nFraction of executed work lost to the modeled interruptions/recovery\n\nfraction\n\nA bounded accounting model multiplies executed work by a quality/acceptance fraction q and the fraction not lost to interruption or recovery.\n\nAssumptions q and $f_{\\mathrm{lost}}$ are measured for the declared task and interval; this is not a universal availability equation.\n\n## Engineering example\n\nA rack can have 64 tested packages yet deliver no useful training step when facility power is uncommissioned, HBM supply is incomplete, or the collective fabric cannot meet the workload’s synchronization bound.\n\n## Distinctions and tradeoffs\n\nA package specification or rack capacity number is not evidence of useful model work without a workload and readiness boundary.\n\nCloser integration reduces some communication distance while increasing assembly, thermal, test and replacement coupling.\n\nLarge models expose the full chain: HBM traffic, collective links, power conversion, cooling, storage and recovery all shape useful progress.\n\n## Specifications and interfaces\n\nClosure boundaries\n\nDie, package, board, rack, hall and workload\n\nEach boundary has its own units and acceptance test.\n\nUseful output\n\nWork retained under stated quality, latency and energy targets\n\nBusy hardware alone is not the output metric.\n\n### Interfaces\n\nDie/package interfaces Input\n\nBumps, PHYs, power and thermal paths\n\nOperating workload Output\n\nMeasured throughput, latency, recovery and quality\n\n## Check your understanding\n\n### What does system closure add after tapeout?\n\n1.  Only a larger logo\n2.  Package, power, cooling, communication, runtime and workload evidence\n3.  Nothing, because a simulated chip is already a data center\n\nAnswer and reasoning\n\nPackage, power, cooling, communication, runtime and workload evidence\n\nUseful computation depends on the complete operating path, not just a logical die.\n\n## Evidence and scope\n\nPublic product and package pages describe representative boundaries; exact fleet configurations, allocation and service telemetry remain configuration-specific.\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-   [TSMC CoWoS technology](https://3dfabric.tsmc.com/english/dedicatedFoundry/technology/cowos.htm)\n-   [NVIDIA DGX GB200 hardware guide](https://docs.nvidia.com/dgx/dgxgb200-user-guide/hardware.html)\n\n[Published lesson data](https://semiaifoundry.com/ai-atlas/lesson-data/crg-package-system-closure-00378de589c06262.json) · [Atlas releases and licensing](https://semiaifoundry.com/ai-atlas/licenses.html)\n\n## Continue exploring\n\n### Foundations\n\n-   [From architecture contract to a chip](https://semiaifoundry.com/learn/atlas/crg-chip-design-realization/)\n-   [Package substrate](https://semiaifoundry.com/learn/atlas/package-substrate/)\n-   [Integrated accelerator rack: GB200 NVL72 example](https://semiaifoundry.com/learn/atlas/rack/)\n\n### Related lessons\n\n-   [GPU logic dies](https://semiaifoundry.com/learn/atlas/gpu-die/)\n-   [High-bandwidth memory stack](https://semiaifoundry.com/learn/atlas/hbm-stack/)\n-   [Distributed checkpoint and recovery state](https://semiaifoundry.com/learn/atlas/checkpoint/)"
}