github-starred/cs249r_book
2
0
Fork 0
mirror of https://github.com/harvard-edge/cs249r_book.git synced 2026-05-06 17:49:07 -05:00
Code Issues 85 Packages Projects Releases 24 Wiki Activity
14,103 Commits 18 Branches 31 Tags
dev
Commit Graph

10 Commits

Author SHA1 Message Date
Vijay Janapa Reddi
285d1e57e6 fix(tutorial): remove ISCA 2026 references — use generic 'Tutorial' label 2026-04-02 15:15:11 -04:00
Vijay Janapa Reddi
0803b97f88 feat(tutorial): upgrade to real beamerthememlsys — branded header, footer, fonts 
Replaced bare Metropolis with the mlsys Beamer theme used by the
textbook lecture slides. Both decks now have:
- Crimson section navigation bar in header
- Branded footer (book logo + Harvard logo + page numbers)
- Helvetica Neue body font + JetBrains Mono/Courier code font
- Section divider pages with progress indicator
- \mlsysfocus for key equation slides

Part 1: 90 pages, 0 errors. Part 2: 47 pages, 0 errors.
 2026-04-02 15:15:11 -04:00
Vijay Janapa Reddi
c02b98d333 fix(tutorial): slides compile cleanly — Part 1 (90pp) + Part 2 (47pp) 
Part 1: fixed \predict→\PredictStart rename, added [fragile] to 13
frames, added xcolor[table] package. Zero errors, 90 pages.

Part 2: added standalone preamble+document wrapper, fixed pareto
figure path, added [fragile] to non-fragile lstlisting frames.
Zero errors, 47 pages.

Total: 137 pages of compiled Beamer slides for ISCA tutorial.
 2026-04-02 15:15:11 -04:00
Vijay Janapa Reddi
c0f46fd04d fix(tutorial): minor Part 2 slide adjustments from simulation feedback 2026-04-02 07:36:16 -04:00
Vijay Janapa Reddi
8928a07ca8 fix(tutorial): simulation round 1 fixes — eta explanation, AllReduce numbers-first, compression fleet 
1. Added 'Understanding eta' slide with table and CPI analogy (before first demo)
2. Added concrete AllReduce example with numbers BEFORE formula
3. Added 'Compression Changes Fleet Architecture' slide with cost table
Addresses top 3 issues from simulation feedback round 1.
 2026-04-02 07:30:43 -04:00
Vijay Janapa Reddi
9955a76b92 feat(staffml): deep verification + mock NeurIPS reviews + paper improvements 
Deep verification: 237-question stratified sample, 4.2% error rate found.
All 10 errors fixed (unit confusion, arithmetic, conceptual misapplication).
96 physics violations removed (impossible topic×track pairs).
Extended invariant checks added (applicability matrix enforcement).

Paper improvements from mock NeurIPS review feedback:
- Bloom critique softened ("complements" not "departs from")
- LLM generation transparency (95% ratio + 4.2% error rate disclosed)
- Scope explicitly limited to technical systems reasoning
- H100 specs corrected (989 TFLOPS, not 495)
- Track percentages reference table instead of hardcoding
- Figure captions use macros for consistency

New topics with questions: software-portability (50), comm-compute-overlap (50).
Phase metadata reclassified (42.5% inference, 37.7% both, 19.9% training).
 2026-04-02 07:28:41 -04:00
Vijay Janapa Reddi
c7bd160e15 feat(tutorial): multi-vendor examples + slide polish for A+ quality 
- Changed A100-vs-H100 comparisons to H100-vs-MI300X-vs-Gaudi3 three-way
- Added Hardware Zoo slide showing all 22 platforms across 6 vendors
- Added timing markers [N min] to speaker notes throughout
- Added "turn to your neighbor" discussion prompts after predict slides
- Added roadmap "You Are Here" slides after each break
- Stripped code boilerplate (import only shown once, aliases for later demos)
- Added figure references for new SVGs (multi-vendor-roofline, parallelism-3d,
  memory-hierarchy, pareto-front)
- Updated related_work.tex comparison table to highlight vendor breadth
 2026-04-02 07:23:12 -04:00
Vijay Janapa Reddi
c7250e32df fix(tutorial): P0 slide blockers — hook, related work, figures, predict, transfer 
1. Added $200M opening hook (2 slides before learning objectives)
2. Wired \input{related_work} between Part 0 and Part 1
3. Replaced figure placeholders with real SVG PDFs (carbon, hardware)
4. Added predict-then-reveal prompts before Aha #3 (256 GPU scaling)
   and Aha #4 (INT4 fleet halving)
5. Added "Name That Wall" personal transfer moment before closing

Addresses all 5 P0 blockers from Patterson's presentation review.
 2026-04-01 23:27:06 -04:00
Vijay Janapa Reddi
0488d01c79 feat(tutorial): add 38 code listings + figure refs to Part 2 slides 
Parts 5-9 were lecture-only (0 code blocks). Now every section has
live demo code using real mlsysim API: EconomicsModel, SustainabilityModel,
DSE search, Engine.sweep, ESP32/nRF52840/H100 comparison, Pipeline
composition, SensitivitySolver, capstone starter code.

Figure placeholders added for carbon-geography and hardware-spectrum SVGs.
 2026-04-01 19:23:31 -04:00
Vijay Janapa Reddi
24ca3b0ef9 feat(mlsysim): complete ISCA tutorial package — backward design + 94 slides 
Full-day tutorial package for ISCA 2026:

Tutorial Design (DESIGN.md, 639 lines):
- Backward design with 4 transfer goals, 6 enduring understandings
- 5 designed "aha moments" with predict-then-reveal structure
- Hour-by-hour schedule with 40% hands-on minimum
- Capstone: multi-region fleet design under budget/latency/carbon constraints
- Facilitation notes, energy management, common pitfalls

Slides (3,081 lines LaTeX):
- Parts 0-4: Welcome, Iron Law, Serving, Compression, Distributed
- Parts 5-9: Economics, DSE, TinyML, Advanced, Wrap-up
- ~94 slides with live demos using real mlsysim API calls
- Speaker notes on every exercise slide

Supporting Materials:
- exercises.md: 8 hands-on exercises with expected answers
- cheatsheet.md: single-page reference card (Iron Law + top equations)
- prerequisites.md: setup instructions + troubleshooting FAQ
 2026-04-01 19:02:23 -04:00