mirror of
https://github.com/harvard-edge/cs249r_book.git
synced 2026-07-19 01:14:07 -05:00
348 lines
12 KiB
YAML
348 lines
12 KiB
YAML
# =============================================================================
|
|
# MLSys·im WEBSITE CONFIGURATION
|
|
# =============================================================================
|
|
# ML Systems Infrastructure Modeling — analytical, not simulation
|
|
# Part of the MLSysBook ecosystem: mlsysbook.ai/mlsysim/
|
|
# =============================================================================
|
|
|
|
project:
|
|
type: website
|
|
output-dir: _build
|
|
resources:
|
|
- mlsysim-paper.pdf
|
|
# Tutorial slide PDFs are injected by mlsysim-publish-live.yml /
|
|
# mlsysim-preview-dev.yml at the site root, alongside the paper.
|
|
- tutorial_module1.pdf
|
|
- tutorial_module2.pdf
|
|
- tutorial_module3.pdf
|
|
- tutorial_module4.pdf
|
|
|
|
website:
|
|
title: "MLSys·im — Machine Learning Systems"
|
|
description: "First-principles infrastructure modeling for ML systems — from edge to exascale."
|
|
site-url: https://mlsysbook.ai/mlsysim/
|
|
favicon: logo.svg
|
|
|
|
open-graph:
|
|
locale: en_US
|
|
site-name: "Machine Learning Systems"
|
|
title: "MLSys·im — ML Systems Infrastructure Modeling"
|
|
description: "Predict ML system performance, cost, and carbon from first principles. Companion to the ML Systems textbook."
|
|
twitter-card:
|
|
title: "MLSys·im — ML Systems Infrastructure Modeling"
|
|
description: "Predict ML performance, cost, and carbon from first principles. Open source."
|
|
card-style: summary
|
|
|
|
page-navigation: true
|
|
reader-mode: false
|
|
back-to-top-navigation: true
|
|
bread-crumbs: true
|
|
|
|
search:
|
|
keyboard-shortcut: ["/"]
|
|
|
|
# Shared navbar from navbar-common.yml; site-local dropdown merged by Quarto.
|
|
# Responsive props (background, pinned, collapse) inherited from
|
|
# shared/config/navbar-common.yml via metadata-files merge.
|
|
navbar:
|
|
logo: assets/images/logo.png
|
|
search: true
|
|
title: "Machine Learning Systems"
|
|
left:
|
|
- text: "MLSys·im"
|
|
menu:
|
|
- icon: calculator
|
|
text: "MLSys·im Home"
|
|
href: ./index.qmd
|
|
- text: "---"
|
|
- icon: play-circle
|
|
text: "Getting Started"
|
|
href: getting-started.qmd
|
|
- icon: terminal
|
|
text: "CLI Reference"
|
|
href: cli-reference.qmd
|
|
- text: "---"
|
|
- icon: journal-code
|
|
text: "Tutorials"
|
|
href: tutorials/index.qmd
|
|
- icon: collection
|
|
text: "The Zoo"
|
|
href: zoo/index.qmd
|
|
- icon: code-square
|
|
text: "API Reference"
|
|
href: api/index.qmd
|
|
- text: "---"
|
|
# Paper lives next to the site it documents, matching the
|
|
# TinyTorch pattern. Path matches the mlsysim-publish-live
|
|
# workflow output (mlsysim-paper.pdf at site root).
|
|
- icon: file-pdf
|
|
text: "MLSys·im Paper (PDF)"
|
|
href: mlsysim-paper.pdf
|
|
target: _blank
|
|
# Tutorial slide PDFs are injected at the site root by the same
|
|
# publish workflow (tutorial_module1.pdf ... tutorial_module4.pdf).
|
|
- icon: easel
|
|
text: "Tutorial Slides"
|
|
href: slides.qmd
|
|
|
|
sidebar:
|
|
style: "floating"
|
|
background: light
|
|
search: true
|
|
collapse-level: 2
|
|
contents:
|
|
# ── GET STARTED ──────────────────────────────────
|
|
- section: "Get Started"
|
|
contents:
|
|
- getting-started.qmd
|
|
- models-and-solvers.qmd
|
|
- cli-reference.qmd
|
|
|
|
# ── LEARN ────────────────────────────────────────
|
|
- section: "Tutorials"
|
|
href: tutorials/index.qmd
|
|
contents:
|
|
- section: "Foundations"
|
|
contents:
|
|
- tutorials/00_hello_roofline.qmd
|
|
- tutorials/01_memory_wall.qmd
|
|
- tutorials/02_two_phases.qmd
|
|
- tutorials/03_kv_cache.qmd
|
|
- section: "Optimization"
|
|
contents:
|
|
- tutorials/04_starving_the_gpu.qmd
|
|
- tutorials/05_quantization.qmd
|
|
- section: "Scale & Cost"
|
|
contents:
|
|
- tutorials/06_scaling_1000_gpus.qmd
|
|
- tutorials/07_geography.qmd
|
|
- tutorials/08_nine_million_dollar.qmd
|
|
- section: "Analysis"
|
|
contents:
|
|
- tutorials/09_sensitivity.qmd
|
|
- tutorials/10_gpu_vs_wafer.qmd
|
|
- tutorials/11_training_memory_capacity_moe.qmd
|
|
- tutorials/12_design_space_exploration.qmd
|
|
- tutorials/12_full_stack_audit.qmd
|
|
- section: "Extending MLSys·im"
|
|
contents:
|
|
- tutorials/01_pipeline_callbacks.qmd
|
|
- tutorials/02_differential_explainer.qmd
|
|
- section: "Self-Paced Tutorial"
|
|
contents:
|
|
- text: "Tutorial Exercises"
|
|
href: tutorial/exercises.qmd
|
|
- text: "Cheat Sheet"
|
|
href: tutorial/cheatsheet.qmd
|
|
- text: "Laws Explained"
|
|
href: laws-explained.qmd
|
|
|
|
# ── INTERACTIVE APPS (future browser exercises) ──
|
|
# - section: "Interactive Apps"
|
|
# contents:
|
|
# - text: "Roofline Explorer"
|
|
# href: interactive_roofline.html
|
|
# - text: "SLA Optimizer"
|
|
# href: sla_optimizer.html
|
|
# - text: "Carbon Placement"
|
|
# href: carbon_placement.html
|
|
|
|
# ── THE ZOO ──────────────────────────────────────
|
|
- section: "The MLSys Zoo"
|
|
href: zoo/index.qmd
|
|
contents:
|
|
- zoo/hardware.qmd
|
|
- zoo/models.qmd
|
|
- zoo/fleets.qmd
|
|
- zoo/infra.qmd
|
|
- zoo/platforms.qmd
|
|
- zoo/datasets.qmd
|
|
- zoo/scenarios.qmd
|
|
- zoo/literature.qmd
|
|
- zoo/ops.qmd
|
|
|
|
# ── FOR YOUR ROLE ────────────────────────────────
|
|
- section: "For Your Role"
|
|
contents:
|
|
- for-students.qmd
|
|
- for-instructors.qmd
|
|
- for-engineers.qmd
|
|
|
|
# ── UNDERSTAND ───────────────────────────────────
|
|
- section: "Concepts"
|
|
contents:
|
|
- philosophy.qmd
|
|
- architecture.qmd
|
|
- math.qmd
|
|
- accuracy.qmd
|
|
- understanding-efficiency.md
|
|
- api-stability.md
|
|
- empirical-calibration.md
|
|
- glossary.qmd
|
|
|
|
# ── BUILD & EXTEND ──────────────────────────────
|
|
- section: "Build & Extend"
|
|
contents:
|
|
- extending-the-engine.qmd
|
|
- provenance.qmd
|
|
- contributing.qmd
|
|
|
|
# ── API REFERENCE ────────────────────────────────
|
|
- section: "API Reference"
|
|
href: api/index.qmd
|
|
contents:
|
|
- text: "Hardware"
|
|
href: api/hardware.qmd
|
|
- text: "Models"
|
|
href: api/models.qmd
|
|
- text: "Systems"
|
|
href: api/systems.qmd
|
|
- text: "Infrastructure"
|
|
href: api/infrastructure.qmd
|
|
- text: "Platforms"
|
|
href: api/platforms.qmd
|
|
- text: "Datasets"
|
|
href: api/datasets.qmd
|
|
- text: "Literature"
|
|
href: api/literature.qmd
|
|
- text: "Ops"
|
|
href: api/ops.qmd
|
|
- text: "Physics"
|
|
href: api/physics.qmd
|
|
- text: "Core"
|
|
href: api/core.qmd
|
|
- text: "Solvers"
|
|
href: api/solvers.qmd
|
|
|
|
# ── BLOG ───────────────────────────────────────────
|
|
- section: "Blog"
|
|
contents:
|
|
- text: "How Much Memory Does Llama-3 70B Need?"
|
|
href: blog/how-much-memory-llama3.md
|
|
|
|
# ── PROJECT ──────────────────────────────────────
|
|
- section: "Project"
|
|
contents:
|
|
- text: "MLSys·im Paper (PDF)"
|
|
href: mlsysim-paper.pdf
|
|
target: _blank
|
|
|
|
# Footer — ecosystem pattern (matches Kits)
|
|
page-footer:
|
|
left: |
|
|
© 2024-2026 Harvard University. Code: <a href="https://www.apache.org/licenses/LICENSE-2.0">Apache-2.0</a> · Docs: <a href="https://creativecommons.org/licenses/by-nc-sa/4.0/">CC-BY-NC-SA 4.0</a>
|
|
center: |
|
|
Part of the <a href="../">Machine Learning Systems</a> textbook
|
|
right:
|
|
- text: |
|
|
<span class="release-pill" data-release-pill></span>
|
|
- icon: github
|
|
href: "https://github.com/harvard-edge/cs249r_book"
|
|
aria-label: "View source on GitHub"
|
|
background: light
|
|
border: true
|
|
|
|
bibliography: references.bib
|
|
csl: config/ieee.csl
|
|
|
|
format:
|
|
html:
|
|
theme:
|
|
light:
|
|
- default
|
|
- styles/style.scss
|
|
dark:
|
|
- default
|
|
- styles/style.scss
|
|
- styles/dark-mode.scss
|
|
respect-user-color-scheme: true
|
|
css: styles/landing.css
|
|
toc: true
|
|
toc-depth: 4
|
|
toc-title: "On this page"
|
|
number-sections: false
|
|
code-copy: true
|
|
code-overflow: wrap
|
|
smooth-scroll: true
|
|
highlight-style: github
|
|
link-external-icon: false
|
|
link-external-newwindow: false
|
|
anchor-sections: true
|
|
fig-width: 7
|
|
fig-height: 4.5
|
|
fig-format: png
|
|
fig-dpi: 120
|
|
include-in-header:
|
|
- file: ../../shared/config/site-head.html
|
|
- text: |
|
|
<meta name="theme-color" content="#0284C7">
|
|
<!-- Release identity (rendered into the footer pill below).
|
|
The publish workflow drops release-manifest.json at the
|
|
site root before deploy. -->
|
|
<meta name="release-manifest" content="/mlsysim/release-manifest.json">
|
|
include-after-body:
|
|
- text: |
|
|
<script src="https://mlsysbook.ai/mlsysim/scripts/subscribe-modal.js" defer></script>
|
|
- file: ../../shared/release/release-pill.html
|
|
|
|
# Announcement bar configuration
|
|
metadata-files:
|
|
- ../../shared/config/navbar-common.yml
|
|
- config/announcement.yml
|
|
|
|
quartodoc:
|
|
package: mlsysim
|
|
dir: api
|
|
title: API Reference
|
|
sections:
|
|
- title: Core API
|
|
desc: Primary objects and resolvers.
|
|
contents:
|
|
- hardware
|
|
- models
|
|
- infrastructure
|
|
- systems
|
|
- platforms
|
|
- datasets
|
|
- literature
|
|
- ops
|
|
- core
|
|
- engine
|
|
- core.provenance.Provenance
|
|
- core.provenance.ProvenanceKind
|
|
- core.provenance.Sourced
|
|
- engine.calibration
|
|
- fmt.fmt
|
|
- fmt.fmt_int
|
|
- physics
|
|
- solvers.SingleNodeModel
|
|
- solvers.NetworkRooflineModel
|
|
- solvers.EfficiencyModel
|
|
- solvers.ForwardModel
|
|
- solvers.ServingModel
|
|
- solvers.TrainingMemoryModel
|
|
- solvers.ServingCapacityModel
|
|
- solvers.ContinuousBatchingModel
|
|
- solvers.WeightStreamingModel
|
|
- solvers.TailLatencyModel
|
|
- solvers.DataModel
|
|
- solvers.TransformationModel
|
|
- solvers.TopologyModel
|
|
- solvers.ScalingModel
|
|
- solvers.InferenceScalingModel
|
|
- solvers.CompressionModel
|
|
- solvers.DistributedModel
|
|
- solvers.MoERoutingModel
|
|
- solvers.ReliabilityModel
|
|
- solvers.OrchestrationModel
|
|
- solvers.EconomicsModel
|
|
- solvers.SustainabilityModel
|
|
- solvers.CheckpointModel
|
|
- solvers.ResponsibleEngineeringModel
|
|
- solvers.SensitivitySolver
|
|
- solvers.SynthesisSolver
|
|
- solvers.ParallelismOptimizer
|
|
- solvers.BatchingOptimizer
|
|
- solvers.PlacementOptimizer
|
|
- engine.dse.DSE
|