Summary: grading & expectations
Active participation (10%): regular participation in class in a curious & positive manner. Unjustified absences will reduce points.
Research proposal presentation (20%): in groups of 2 students you will submit and present a short one-page research proposal related to sustainable computing, graded by your fellow students.
Individual paper presentation (70%): present one full conference paper/two short workshop papers.
Prerequisites: Operating Systems (234123).Further knowledge in operating systems and computer architecture will help to understand the topics presented.
Active participation
You are expected to regularly raise an idea or questions that demonstrate interest in the topic and shown results, constructive thinking on potential follow-ups, etc.
Individual paper presentation
Each student gives one individual presentation (one class = one student) on one full paper (or two short papers, ~22 min each).
Time frame:
• Full paper: 40–45 min presentation + 5 min questions
• Two short papers: 22 min per paper + 3 min questions per-paper
What I expect from presenters
• Try to enjoy. Boredom is the enemy of learning
• Arrive early, set up, start on time, finish on time. Rehearse multiple times to ensure you’re on time
• Explain paper’s claims (consider stating the innovation highlights early on)
• Explain required background (if necessary)
• Clearly explain each important figure (setup, axes, whether higher is better, takeaways)
• If equations are necessary: slowly walk through them
• Highlight paper’s strengths, weaknesses, and open questions— apply critical thinking, e.g., claims from industry carry different weight
• Engage the audience occassionally to keep people listening (e.g., questions, short checks, visuals)
• Practice beforehand (ideally in front of someone)
• Golden rule: aim for ~4 one-line bullets per slide; avoid dense text that no one reads
You won’t be penalized for minor speech stumbles or using notes to alleviate stress; points will be reduced for not rehearsing, significantly running over time, failing to explain the paper’s claims/figures, and copy-pasting paper text to slides.
Research Proposal Presentation
Teams of two students will give an 8–10 minute presentation, proposing a new research idea related to sustainability in computing. Presentations will be scheduled in the final classes of the semester.
Grading is based on peer evaluations, but topics must be approved by me.
Choosing a Topic: Any original idea with a plausible,potentially measurable decarbonization impact is welcome. For inspiration, browse recent work at HotCarbon (https://hotcarbon.org/).
Graduate/PhD students are encouraged to choose a topic aligned with their research.
Use Google Scholar, LLMs, etc., to help you understand whether someone has already done what you're proposing!
Proposal document format (one page, font 11, in English only)
• Names + student IDs
• Title
• Background & related work (1–2 paragraphs)
• Your proposed solution + what makes it innovative
• Methods (1-2 paragraph): planned approach, methodology,experiments, and evaluation
• Potential sustainability impact (1 paragraph): how your idea improves sustainability in computing
• Related work: (1-2 paragraphs): previous research or commercial solutions most similar or closely related to your work
• Optional appendix: one page for references + figures
• Mandatory appendix: brief summary of three main criticisms from a known LLM (of your choosing) and how you respond to them, using a critical reviewer prompt, i.e.., "You are a highly critical veteran in the field of computer science academic research. See attached a research proposal for an academic paper in computer science on the topic of sustainability in computing. Do you think this idea is interesting and could be published in a top-tier conference?explain your reasoning".
Presentation: based on the proposal. Be sure to include the main issues raised by the virtual critique and explain how you address them (or intend to)
Grading
• Mostly based on evaluation forms filled out by your peers (after each presentation)
• Grade determined according to summary of forms
• Anonymous input will be provided based on the forms
Important dates:
Exact dates will be published later. Expect to submit proposal around June, and present in 3 final classes
Papers & assignments
15.4: introduction
Modelling:
1. [אוהד איתן 29.4] ACT: Designing Sustainable Computer Systems With An Architectural Carbon Modeling Tool, ISCA'22
2. [שירז שמולמן 29.4] FOCAL: A First-Order Carbon Model to Assess Processor Sustainability, ASPLOS'24
22.4 יום העצמאות
3. [6.5 מישל מנדלייל] (short) CarbonClarity: Understanding and Addressing Uncertainty in Embodied Carbon for Sustainable Computing, arxiv 2025
+
(short) On the Promise and Pitfalls of Optimizing Embodied Carbon, EIR'24
Cloud
4. [6.5 ויקטור אברבוך] Carbon accounting in the Cloud (Google), 2024
5. [13.5 נדב בלוגר] Fair-CO2: Fair Attribution for Cloud Carbon Emissions, ISCA'25
6. [13.5 יניב קניאל] (short) On the Implications of Choosing Average versus Marginal Carbon Intensity Signals on Carbon-aware Optimizations , E-Energy'24
+
(short) The Sunk Carbon Fallacy: Rethinking Carbon Footprint Metrics for Effective Carbon-Aware Scheduling, SoCC'24
Cooling water management:
7. [20.5 טל סרוסי] ThirstyFLOPS:Water Footprint Modeling and Analysis Toward Sustainable HPC Systems, SC'25
8. [20.5 לין כבהא] WaterWise: Co-optimizing Carbon- and Water-Footprint Toward Environmentally Sustainable Cloud Computing, PPoPP'25
AI:
9. [20.5 מוחמד עסאף] Toward Sustainable HPC: Carbon Footprint Estimation and Environmental Implications of HPC Systems, SC'25
10. [27.5 שקד גולדשטיין] GREEN: Carbon-efficient Resource Scheduling for Machine Learning Clusters, NSDI'25
11. [27.5 שקד עדי] CarbonEdge: Leveraging Mesoscale Spatial Carbon-Intensity Variations for Low Carbon Edge Computing, HPDC'25
12 [3.6 אמיר סורני] (short) Measuring the environmental impact of delivering AI at Google Scale, 2025
+
(short) From Efficiency Gains to Rebound Effects: The Problem of Jevons’ Paradox in AI’s Polarized Environmental Debate, Faact'25
Storage:
13. [3.6 איתמר קיגיס] Towards Efficient Flash Caches with Emerging NVMe Flexible Data Placement SSDs, Eurosys'25
14. [10.6 אופיר מזרחי] (Short) Embodied Carbon Footprint of 3D NAND Memories, CF'2025
+
(short) Carbon-Aware Memory Placement, EIR'24
Applications:
15. [10.6 מיכאל בלום] Environmental Footprints of Query Processing: A Vision for Sustainable Database Architectures, VLDB'25
16. [17.6 יועד אייבלס] Systematic Analysis of Kernel Security Performance and Energy Costs, ASIA CCS'25
17. [17.6 לנא אבו אל היגא] GreenThrift: Optimizing Carbon and Cost for Flexible Residential Loads, JCSS'25
Web:
18. [24.6 רון וליצקי] CDN-Shifter: Leveraging SpatialWorkload Shifting to Decarbonize Content Delivery Networks, SoCC'24
19. [24.6 אימאן ברבור] Slower is Greener: Acceptance of Eco-feedback Interventions on Carbon Heavy Internet Services, JCSS'25
+ הצגת הצעות מחקר