Evaluating Computing Platforms for Sustainability: A Comparative Analysis of FPGAs against ASICs, GPUs, and CPUs

Chetan Choppali Sudarshan, Aman Arora, Vidya A. Chhabria · 2026

Read Evaluating Computing Platforms for Sustainability: A Comparative Analysis of FPGAs against ASICs, GPUs, and CPUs by Chetan Choppali Sudarshan, Aman Arora, Vidya A. Chhabria (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-20638.txt.

Introduction

Sudarshan, Arora, and Chhabria offer GreenFPGA, a tool that gauges the carbon footprint of FPGAs over their life with uncertainty. Carbon footprint covers running and built-in gases from use, making, and design. FPGAs stand out through reuse, spreading built-in footprint across many uses. GreenFPGA counts footprint from design, making, reuse, use, scrap, testing, and recycling, and finds regimes where FPGAs beat ASICs, GPUs, and CPUs at equal output. Across equal-output case studies, FPGAs prove more green for often changing, diverse workloads and low-volume uses.

Catalog record

CREATORChetan Choppali Sudarshan, Aman Arora, Vidya A. Chhabria
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE78013
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.20638

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Evaluating Computing Platforms for Sustainability: A Comparative Analysis of FPGAs against ASICs, GPUs, and CPUs Chetan Choppali Sudarshan, Aman Arora, Vidya A. Chhabria Abstract . Climate change concerns emphasize the need for sustainable computing. Modeling the carbon footprint (CFP), including operational and embodied CFP from semiconductor use, manufacture and design, is essential. Field programmable gate arrays (FPGAs) stand out as promising platforms due to their reconfigurability across various applications, enabling the amortization of embodied CFP across multiple applications. This paper introduces GreenFPGA, a tool estimating the total CFP of FPGAs over their lifespan, considering uncertainties in CFP modeling. It accounts for CFP during design, manufacturing, reconfigurability (reuse), operation, disposal, testing, and recycling. GreenFPGA identifies deployment regimes in which FPGAs can be more sustainable than ASICs, GPUs, and CPUs under the modeled iso-performance assumptions. Experimental results highlight the importance of analyzing applications across different computing platforms to assess their CFP while varying parameters such as application type, lifetime, usage time, and volume impact their total CFP. Across the evaluated pairwise iso-performance case studies with ASICs, GPUs, and CPUs, FPGAs can be more sustainable under specific deployment regimes involving frequently changing, diverse workloads and low-volume applications. 1. Introduction From microchips to data centers, computing carries a significant carbon footprint (CFP). Rising computational demand, fueled by applications such as artificial intelligence (AI) ( 27 ) , has made the information and communication technology (ICT) sector responsible for 2.1%–3.9% of global CFP ( 19 ) . While the semiconductor industry has focused on making chips smaller, faster, and more energy-efficient to reduce operational CFP, the ecological impact of design, manufacturing, and disposal—embodied CFP—is equally critical ( 24 ) . Previous studies ( 24 ; 48 ) highlight the significant impact of embodied carbon in data centers and edge devices, offering tools to calculate CFP for monolithic and chiplet-based systems. However, they overlook field-programmable gate arrays (FPGAs) and reconfigurable platforms. While ECO-CHIP ( 48 ; 46 ) applies the 3R concept (reduce-reuse-recycle) by reusing chiplets, this paper introduces a fourth “R,” focusing on reconfigurable computing to further reduce CFP ( 11 ) . When considering the spectrum of processing platforms, including FPGAs, graphics processing units (GPUs), central processing units (CPUs), and application-specific integrated circuits (ASICs), FPGAs possess a high degree of reconfigurability, allowing developers to tailor hardware to the demands of specific applications precisely. CPUs, while inherently more general-purpose than GPUs, retain a degree of flexibility that enables their deployment across a broader range of tasks. However, the differing levels of flexibility seen in GPUs, CPUs, and FPGAs come at the cost of varying energy efficiency, making some platforms less efficient than ASICs. Analyzing the full lifecycle CFP of these devices from a sustainability standpoint is crucial for identifying the most suitable computing platform for an application. Figure 1 . Lifecycle of a computing platform: Highlighting the detailed embodied and operational CFP from cradle to grave. Fig. 1 shows the complete lifecycle of a computing platform or a chip. It outlines the factors contributing to embodied CFP, including design, manufacturing, memory, package, testing, discard, retire, and end-of-life (EOL), and operational CFP tied to end-user activities, reconfigurability, and reliability. Here, retire denotes the terminal handling of the device through discard or recycle, and EOL refers to replacement-related lifecycle effects caused by failure, obsolescence, or upgrade-driven turnover. The key difference compared to ASICs is that FPGAs, CPUs, and GPUs can be reconfigured or reprogrammed to be used across multiple applications, as shown in the green box, while ASICs are at the end of life once the application’s lifetime is complete. In this paper, we introduce GreenFPGA, a tool to assess the CFP of FPGA-based computing throughout its lifecycle shown in Fig. 1 . We focus on evaluating the CFP of FPGAs, and develop new models that can account for these FPGA-related distinctions. While FPGAs exhibit larger physical size and lower energy efficiency compared to equivalent ASICs, leading to increased embodied and operational CFP, they offer compelling sustainability advantages: • Reconfigurability: FPGAs are reconfigurable and adaptable for multiple applications post-manufacturing, a unique feature absent in ASICs and limited in GPUs. • Extended lifespan: FPGAs typically have longer lifespans and use cases lasting 12 to 15 years ( 2 ) as they can be reconfigured, compared to ASICs and GPUs that become obsolete with rapidly changing applications (2 to 8 years). • Energy-efficient: FPGAs can offer higher energy-efficiency compared to CPUs and GPUs in some scenarios, as they can be programmed and tailored at the hardware level to meet specific task requirements, minimizing unnecessary operations and energy usage. Over time and across diverse applications, unlike ASICs, the embodied CFP incurred during the manufacturing and design of FPGAs, GPUs, and CPUs can be amortized over their lifespans and across their uses. Therefore, they make a compelling case from a sustainability perspective for analyzing the CFP. Using GreenFPGA, we analyze scenarios to determine when FPGAs are more sustainable alternatives for computing than ASICs, CPUs, and GPUs. For instance, Fig. 3 compares FPGAs and ASICs under conditions where they have similar performance (iso-performance) for a given application (iso-performance). The line plot suggests FPGAs are more sustainable for use cases with more than three applications (similar experimental setup as Sec. 5.1 ). Similarly, Fig. 3 compares ASICs with FPGA at iso-performance ( 49 ) (similar experiment setup as Sec. 5.1 ). When the number of applications (NumApp) is equal to 1, FPGAs have higher CFP compared to ASICs (+164%) as they have larger areas and lower energy efficiencies for these testcases. However, when NumApp = 5, FPGAs exhibit a lower CFP (-46%) than ASICs due to amortization of the embodied CFP. We build GreenFPGA, a CFP modeling tool for FPGAs, and compare them against CPUs, ASICs, and GPUs. The key contributions of our work are as follows. Figure 2 . Comparing the CFP of FPGAs vs. ASICs at iso-performance while sweeping the number of applications using a probabilistic model (box plot data), and a deterministic model ( 13 ) (line plot). Figure 3 . CFP comparison between ASIC and FPGA-based computing for a single application and five applications. (1) To the best of our knowledge, our work, GreenFPGA, is the first to model and assess the CFP of FPGA across its entire lifespan by accounting for the unique aspects of FPGA-based computing, including CFP overheads from reconfiguring the FPGA and application development time. (2) Unlike prior work in CFP modeling of chips ( 48 ; 24 ; 56 ; 13 ; 57 ) , GreenFPGA models replacement effects (EOL of devices) due to device failure, obsolescence, refresh/upgrade, and is the first to incorporate reliability-aware degradation of energy efficiency for operational carbon. (3) We develop a more robust model for carbon modeling of the design phase of a chip’s lifecycle compared to ( 48 ) based on industry (fabless design houses) reports, and also develop a model for testing-related embodied CFP. (4) Given that RAM can contribute significantly up to 38% ( 41 ) to the overall total CFP, our framework now includes the embodied CFP of RAM and not just the processor as in prior works ( 13 ; 24 ) . (5) We extend the deterministic model in ( 13 ) to

Read the complete text