The rapid increase in computing demand and corresponding energy consumption have focused attention on computing’s impact on the climate and sustainability. Prior work proposes metrics that quantify computing’s carbon footprint across several lifecycle phases, including its supply chain, operation, and end-of-life. Industry uses these metrics to optimize the carbon footprint of manufacturing hardware and running computing applications. Unfortunately, prior work on optimizing datacenters’ carbon footprint often succumbs to the sunk cost fallacy by considering embodied carbon emissions (a sunk cost) when making operational decisions (i.e., job scheduling and placement), which leads to operational decisions that do not always reduce the total carbon footprint.In this paper, we evaluate carbon-aware job scheduling and placement on a given set of servers for several carbon accounting metrics. Our analysis reveals state-of-the-art carbon accounting metrics that include embodied carbon emissions when making operational decisions can increase the total carbon footprint of executing a set of jobs. We study the factors that affect the added carbon cost of such suboptimal decision-making. We then use a real-world case study from a datacenter to demonstrate how the sunk carbon fallacy manifests itself in practice. Finally, we discuss the implications of our findings in better guiding effective carbon-aware scheduling in on-premise and cloud datacenters.
@inproceedings{10.1145/3698038.3698542,author={Bashir, Noman and Gohil, Varun and Subramanya, Anagha Belavadi and Shahrad, Mohammad and Irwin, David and Olivetti, Elsa and Delimitrou, Christina},title={The Sunk Carbon Fallacy: Rethinking Carbon Footprint Metrics for Effective Carbon-Aware Scheduling},year={2024},isbn={9798400712869},publisher={Association for Computing Machinery},address={New York, NY, USA},url={https://doi.org/10.1145/3698038.3698542},doi={10.1145/3698038.3698542},booktitle={Proceedings of the 2024 ACM Symposium on Cloud Computing},pages={542–551},numpages={10},keywords={Sustainable computing, datacenters, metrics, operational and embodied carbon footprint, scheduling, sustainability},location={Redmond, WA, USA},series={SoCC '24},}
The Importance of Generalizability in Machine Learning for Systems
@article{genmlforsys,author={Gohil, Varun and Dev, Sundar and Upasani, Gaurang and Lo, David and Ranganathan, Parthasarathy and Delimitrou, Christina},journal={IEEE Computer Architecture Letters},title={The Importance of Generalizability in Machine Learning for Systems},year={2024},volume={23},number={01},issn={1556-6064},pages={95-98},keywords={bandwidth;data models;bayes methods;uncertainty;computational modeling;predictive models;internet},doi={10.1109/LCA.2024.3384449},publisher={IEEE Computer Society},address={Los Alamitos, CA, USA},month=jan,}
Sabre: Hardware-Accelerated Snapshot Compression for Serverless MicroVMs
@inproceedings{sabre,author={Lazarev, Nikita and Gohil, Varun and Tsai, James and Anderson, Andy and Chitlur, Bhushan and Zhang, Zhiru and Delimitrou, Christina},title={Sabre: {Hardware-Accelerated} Snapshot Compression for Serverless {MicroVMs}},booktitle={18th USENIX Symposium on Operating Systems Design and Implementation (OSDI 24)},year={2024},isbn={978-1-939133-40-3},address={Santa Clara, CA},pages={1--18},url={https://www.usenix.org/conference/osdi24/presentation/lazarev},publisher={USENIX Association},month=jul,video={https://www.usenix.org/conference/osdi24/presentation/lazarev},}
2021
Performance optimization opportunities in the Android software stack
@inproceedings{GOHIL2021100003,title={Performance optimization opportunities in the Android software stack},booktitle={BenchCouncil Transactions on Benchmarks Standards},volume={1},number={1},pages={100003},year={2021},issn={2772-4859},doi={https://doi.org/10.1016/j.tbench.2021.100003},url={https://www.sciencedirect.com/science/article/pii/S277248592100003X},author={Gohil, Varun and Ujjainkar, Nisarg and Mekie, Joycee and Awasthi, Manu},keywords={CPU utilization, Android smartphone, Workload characterization},video={https://youtu.be/rXfwBpVlTi0},}
Fixed-Posit: A Floating-Point Representation for Error-Resilient Applications
@inproceedings{9399648,author={Gohil, Varun and Walia, Sumit and Mekie, Joycee and Awasthi, Manu},booktitle={IEEE Transactions on Circuits and Systems II: Express Briefs},title={Fixed-Posit: A Floating-Point Representation for Error-Resilient Applications},year={2021},volume={68},number={10},pages={3341-3345},doi={10.1109/TCSII.2021.3072217},}
2020
[Re] One ticket to win them all: generalizing lottery ticket initializations across datasets and optimizers
@article{gohil2019one,title={[Re] One ticket to win them all: generalizing lottery ticket initializations across datasets and optimizers},author={Gohil, Varun and Narayanan, S Deepak and Jain, Atishay},journal={NeurIPS Reproducibility Challenge 2019, ReScience-C},year={2020},}
@inproceedings{10.1145/3371158.3371230,author={Dutta, Ritik and Gohil, Varun and Jain, Atishay},title={Effect of Feature Hashing on Fair Classification},year={2020},isbn={9781450377386},publisher={Association for Computing Machinery},address={New York, NY, USA},url={https://doi.org/10.1145/3371158.3371230},doi={10.1145/3371158.3371230},booktitle={Young Researchers Symposium, Proceedings of the 7th ACM IKDD CoDS and 25th COMAD},pages={365–366},numpages={2},location={Hyderabad, India},series={CoDS COMAD 2020},}
@inproceedings{254296,author={V, Subisha and Gohil, Varun and Ujjainkar, Nisarg and Awasthi, Manu},title={Prefetching in Hybrid Main Memory Systems},booktitle={12th {USENIX} Workshop on Hot Topics in Storage and File Systems (HotStorage 20)},year={2020},url={https://www.usenix.org/conference/hotstorage20/presentation/v},publisher={{USENIX} Association},month=jul,video={https://2459d6dc103cb5933875-c0245c5c937c5dedcca3f1764ecc9b2f.ssl.cf2.rackcdn.com/hotstorage20/hotstorage20-paper86-video-long-gohil.mp4},}
@inproceedings{Gohil:2019:FFA:3302541.3313102,author={Gohil, Varun and Singh, Shreyas and Awasthi, Manu},title={FAB: Framework for Analyzing Benchmarks},booktitle={Companion of the 2019 ACM/SPEC International Conference on Performance Engineering},series={ICPE '19},year={2019},isbn={978-1-4503-6286-3},location={Mumbai, India},pages={33--36},numpages={4},url={http://doi.acm.org/10.1145/3302541.3313102},doi={10.1145/3302541.3313102},acmid={3313102},publisher={ACM},address={New York, NY, USA},keywords={instruction mix, workload analysis, workload selection, workload similarity},}
@inproceedings{Parikh:2018:MME:3241539.3267737,author={Parikh, Nisarg and Gohil, Varun and Awasthi, Manu},title={META: Memory Exploration Tool for Android Devices},booktitle={Poster Track, Proceedings of the 24th Annual International Conference on Mobile Computing and Networking},series={MobiCom '18},year={2018},isbn={978-1-4503-5903-0},location={New Delhi, India},pages={774--776},numpages={3},url={http://doi.acm.org/10.1145/3241539.3267737},doi={10.1145/3241539.3267737},acmid={3267737},publisher={ACM},address={New York, NY, USA},keywords={memory hierarchy, mobile computer architecture, simulation},}