Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Core Hours and Carbon Credits: Incentivizing Sustainability in HPC

Alok Kamatar, Maxime Gonthier, Valerie Hayot-Sasson, Andre Bauer, Marcin Copik, Torsten Hoefler, Raul Castro Fernandez, Kyle Chard, Ian Foster

TL;DR

Energy and carbon emissions in HPC are not adequately priced to drive sustainability. The authors propose Energy- and Carbon-Based Accounting (EBA/CBA) and implement Green-ACCESS, a FaaS-based platform enabling fungible, environmental-impact–driven allocations across CPUs and GPUs. They validate the approach with a large-scale user survey, hardware experiments, simulations, and a user study, showing that environmental pricing can steer users toward energy-efficient resources and increase effective work per allocation. The work provides concrete mechanisms and a deployable prototype to align incentives for reduced energy use and carbon emissions in HPC settings.

Abstract

Realizing a shared responsibility between providers and consumers is critical to manage the sustainability of HPC. However, while cost may motivate efficiency improvements by infrastructure operators, broader progress is impeded by a lack of user incentives. We conduct a survey of HPC users that reveals fewer than 30 percent are aware of their energy consumption, and that energy efficiency is among users' lowest priority concerns. One explanation is that existing pricing models may encourage users to prioritize performance over energy efficiency. We propose two transparent multi-resource pricing schemes, Energy- and Carbon-Based Accounting, that seek to change this paradigm by incentivizing more efficient user behavior. These two schemes charge for computations based on their energy consumption or carbon footprint, respectively, rewarding users who leverage efficient hardware and software. We evaluate these two pricing schemes via simulation, in a prototype, and a user study.

Core Hours and Carbon Credits: Incentivizing Sustainability in HPC

TL;DR

Energy and carbon emissions in HPC are not adequately priced to drive sustainability. The authors propose Energy- and Carbon-Based Accounting (EBA/CBA) and implement Green-ACCESS, a FaaS-based platform enabling fungible, environmental-impact–driven allocations across CPUs and GPUs. They validate the approach with a large-scale user survey, hardware experiments, simulations, and a user study, showing that environmental pricing can steer users toward energy-efficient resources and increase effective work per allocation. The work provides concrete mechanisms and a deployable prototype to align incentives for reduced energy use and carbon emissions in HPC settings.

Abstract

Realizing a shared responsibility between providers and consumers is critical to manage the sustainability of HPC. However, while cost may motivate efficiency improvements by infrastructure operators, broader progress is impeded by a lack of user incentives. We conduct a survey of HPC users that reveals fewer than 30 percent are aware of their energy consumption, and that energy efficiency is among users' lowest priority concerns. One explanation is that existing pricing models may encourage users to prioritize performance over energy efficiency. We propose two transparent multi-resource pricing schemes, Energy- and Carbon-Based Accounting, that seek to change this paradigm by incentivizing more efficient user behavior. These two schemes charge for computations based on their energy consumption or carbon footprint, respectively, rewarding users who leverage efficient hardware and software. We evaluate these two pricing schemes via simulation, in a prototype, and a user study.
Paper Structure (29 sections, 5 equations, 10 figures, 6 tables)

This paper contains 29 sections, 5 equations, 10 figures, 6 tables.

Table of Contents

  1. Introduction
  2. Survey
  3. Survey Design
  4. Results
  5. Impact-Based Accounting
  6. Background: Fungible Allocations
  7. Energy-Based Accounting Model
  8. Carbon-Based Accounting Model
  9. Green-ACCESS System Implementation
  10. Green-ACCESS Design
  11. Comparison of Accounting Methods
  12. Green-ACCESS on CPUs
  13. Green-ACCESS on GPUs
  14. Comparison of Embodied Carbon
  15. Simulation Studies
  16. ...and 14 more sections

Figures (10)

  • Figure 1: Responses to question "Are you aware of how the HPC resources you use perform on the following sustainability metrics?"
  • Figure 2: Energy efficiency is among the least important factors for users when choosing where to run a job.
  • Figure 3: Architecture of the green-ACCESS prototype. The three principal system components, shaded green, grey, and orange, are described in the text.
  • Figure 4: Runtime and energy consumption of seven applications running on four different nodes reveal different performance/energy tradeoffs across machines.
  • Figure 5: Results from simulating EBA. The Energy and Greedy policies complete the most work by prioritizing FASTER over IC for efficiency. The Mixed policy trades some efficiency for faster completion time.
  • ...and 5 more figures