{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,1,23]],"date-time":"2026-01-23T05:39:44Z","timestamp":1769146784357,"version":"3.49.0"},"publisher-location":"New York, NY, USA","reference-count":35,"publisher":"ACM","content-domain":{"domain":["dl.acm.org"],"crossmark-restriction":true},"short-container-title":[],"published-print":{"date-parts":[[2026,1,26]]},"DOI":"10.1145\/3784828.3785157","type":"proceedings-article","created":{"date-parts":[[2026,1,22]],"date-time":"2026-01-22T13:19:17Z","timestamp":1769087957000},"page":"31-39","update-policy":"https:\/\/doi.org\/10.1145\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["System-Level Energy Profiling of Wafer-Scale AI Systems: Characterizing Non-Accelerator Overheads in the Cerebras CS-2 System"],"prefix":"10.1145","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-6670-5343","authenticated-orcid":false,"given":"Jophin","family":"John","sequence":"first","affiliation":[{"name":"Leibniz Supercomputing Centre (LRZ), Garching near Munich, Germany"}]},{"ORCID":"https:\/\/orcid.org\/0009-0000-4049-2204","authenticated-orcid":false,"given":"Hoi-Fong","family":"Mak","sequence":"additional","affiliation":[{"name":"Leibniz Supercomputing Centre (LRZ), Garching near Munich, Germany"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-5003-5138","authenticated-orcid":false,"given":"Michael","family":"Hoffmann","sequence":"additional","affiliation":[{"name":"Leibniz Supercomputing Centre (LRZ), Garching near Munich, Germany"}]},{"ORCID":"https:\/\/orcid.org\/0009-0009-9845-8557","authenticated-orcid":false,"given":"Alice","family":"Zhang","sequence":"additional","affiliation":[{"name":"Cerebras Systems, Sunnyvale, USA"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2543-9688","authenticated-orcid":false,"given":"Tapasya","family":"Patki","sequence":"additional","affiliation":[{"name":"Lawrence Livermore National Laboratory, Livermore, USA"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-6258-8813","authenticated-orcid":false,"given":"Nicolay","family":"Hammer","sequence":"additional","affiliation":[{"name":"Leibniz Supercomputing Centre (LRZ), Garching near Munich, Germany"}]}],"member":"320","published-online":{"date-parts":[[2026,1,25]]},"reference":[{"key":"e_1_3_3_1_2_2","volume-title":"Energy and AI, IEA","author":"(2025) IEA","year":"2025","unstructured":"IEA (2025). 2025. Energy and AI, IEA. ttps:\/\/www.iea.org\/reports\/energy-and-ai"},{"key":"e_1_3_3_1_3_2","doi-asserted-by":"publisher","DOI":"10.1109\/SC.2014.18"},{"key":"e_1_3_3_1_4_2","unstructured":"American Public Power Association. 2024. AI to Drive 165% Increase in Data Center Power Demand by 2030: Goldman Sachs. American Public Power Association (2024). https:\/\/www.publicpower.org\/periodical\/article\/ai-drive-165-increase-data-center-power-demand-2030-goldman-sachs Accessed: 2025-07-25."},{"key":"e_1_3_3_1_5_2","unstructured":"Cerebras Systems. 2022. Scaling Up and Out: Training Massive Models on Cerebras Systems using Weight Streaming. Cerebras Blog (2022). https:\/\/www.cerebras.ai\/blog\/scaling-up-and-out-training-massive-models-on-cerebras-systems-using-weight-streaming Accessed: 2025-07-25."},{"key":"e_1_3_3_1_6_2","unstructured":"Cerebras Systems. 2024. Cerebras CS-3. Cerebras Blog (2024). https:\/\/www.cerebras.ai\/blog\/cerebras-cs3 Accessed: 2025-07-25."},{"key":"e_1_3_3_1_7_2","unstructured":"Cerebras Systems. 2024. Documentation for Developing with CSL. https:\/\/sdk.cerebras.net\/. Accessed: 2025-12-15."},{"key":"e_1_3_3_1_8_2","unstructured":"Cerebras Systems. 2024. Get Started with Cerebras. https:\/\/training-docs.cerebras.ai\/. Accessed: 2025-12-15."},{"key":"e_1_3_3_1_9_2","volume-title":"WSE-2 Data Sheet","author":"Inc. Cerebras Systems","year":"2021","unstructured":"Cerebras Systems Inc.2021. WSE-2 Data Sheet. Technical Report. Cerebras Systems Inc.https:\/\/f.hubspotusercontent30.net\/hubfs\/8968533\/WSE-2%20Datasheet.pdf"},{"key":"e_1_3_3_1_10_2","unstructured":"Mohak Chadha. [n. d.]. Adaptive Resource-Aware Batch Scheduling for HPC systems. ([n. d.])."},{"key":"e_1_3_3_1_11_2","first-page":"1255","volume-title":"INFORMATIK 2023 - Designing Futures: Zuk\u00fcnfte gestalten","author":"Dokic Dusan","year":"2023","unstructured":"Dusan Dokic, Hannah Stein, Sabine Janzen, and Wolfgang Maa\u00df. 2023. Towards Energy-Efficient Large-Scale Artificial Intelligence for Sustainable Data Centers. In INFORMATIK 2023 - Designing Futures: Zuk\u00fcnfte gestalten. Gesellschaft f\u00fcr Informatik e.V., Bonn, 1255\u20131265. https:\/\/doi.org\/10.18420\/inf2023134"},{"key":"e_1_3_3_1_12_2","doi-asserted-by":"publisher","unstructured":"Alicia Golden Mariam Elgamal Abdulrahman Mahmoud Gage Hills Carole-Jean Wu Gu-Yeon Wei and David Brooks. 2025. Wafer-Scale Systems: A Carbon Perspective. SIGENERGY Energy Inform. Rev. 5 2 (Aug. 2025) 118\u2013124. 10.1145\/3757892.3757909","DOI":"10.1145\/3757892.3757909"},{"key":"e_1_3_3_1_13_2","unstructured":"Aaron Grattafiori and et al.2024. The Llama 3 Herd of Models. arxiv:https:\/\/arXiv.org\/abs\/2407.21783\u00a0[cs.AI] https:\/\/arxiv.org\/abs\/2407.21783"},{"key":"e_1_3_3_1_14_2","doi-asserted-by":"publisher","DOI":"10.3233\/FAIA250929"},{"key":"e_1_3_3_1_15_2","doi-asserted-by":"publisher","DOI":"10.1109\/ICTAI62512.2024.00048"},{"key":"e_1_3_3_1_16_2","unstructured":"Michael Hoffmann Jophin John Stefan Schweter Gokul Ramakrishnan Hoi-Fong Mak Alice Zhang Dmitry Gaynullin and Nicolay\u00a0J. Hammer. 2025. Llama-GENBA-10B: A Trilingual Large Language Model for German English and Bavarian. arxiv:https:\/\/arXiv.org\/abs\/2509.05668\u00a0[cs.CL] https:\/\/arxiv.org\/abs\/2509.05668"},{"key":"e_1_3_3_1_17_2","unstructured":"Edward\u00a0J. Hu Yelong Shen Phillip Wallis Zeyuan Allen-Zhu Yuanzhi Li Shean Wang Lu Wang and Weizhu Chen. 2021. LoRA: Low-Rank Adaptation of Large Language Models. arxiv:https:\/\/arXiv.org\/abs\/2106.09685\u00a0[cs.CL] https:\/\/arxiv.org\/abs\/2106.09685"},{"key":"e_1_3_3_1_18_2","unstructured":"Integrity Energy. 2024. Why is Data Center Energy Consumption So High? Integrity Energy Blog (2024). https:\/\/www.integrityenergy.com\/blog\/why-is-data-center-energy-consumption-so-high\/ Accessed: 2025-07-25."},{"key":"e_1_3_3_1_19_2","unstructured":"Intel Corporation. 2024. Intel\u00ae VTune\u2122 Profiler. Intel Developer Tools. https:\/\/www.intel.com\/content\/www\/us\/en\/developer\/tools\/oneapi\/vtune-profiler.html Accessed: 2025-07-25."},{"key":"e_1_3_3_1_20_2","unstructured":"Jophin John. 2024. Leveraging Dynamic Resource Management for Power Management and Fault Tolerance in High Performance Computing. Ph.\u00a0D. Dissertation. Technische Universit\u00e4t M\u00fcnchen. https:\/\/mediatum.ub.tum.de\/1739065"},{"key":"e_1_3_3_1_21_2","doi-asserted-by":"crossref","unstructured":"Jophin John Santiago Narv\u00e1ez and Michael Gerndt. 2020. Invasive computing for power corridor management. Parallel Computing: Technology Trends 36 (2020) 386.","DOI":"10.3233\/APC200063"},{"key":"e_1_3_3_1_22_2","unstructured":"Justin Selig. [n. d.]. The Cerebras Software Development Kit: A Technical Overview. https:\/\/8968533.fs1.hubspotusercontent-na2.net\/hubfs\/8968533\/Cerebras%20SDK%20Technical%20Overview%20White%20Paper.pdf"},{"key":"e_1_3_3_1_23_2","doi-asserted-by":"publisher","unstructured":"Sean Lie. 2023. Cerebras Architecture Deep Dive: First Look Inside the Hardware\/Software Co-Design for Deep Learning. IEEE Micro 43 3 (2023) 18\u201330. 10.1109\/MM.2023.3256384","DOI":"10.1109\/MM.2023.3256384"},{"key":"e_1_3_3_1_24_2","unstructured":"Haotian Liu Chunyuan Li Qingyang Wu and Yong\u00a0Jae Lee. 2023. Visual Instruction Tuning. arxiv:https:\/\/arXiv.org\/abs\/2304.08485\u00a0[cs.CV] https:\/\/arxiv.org\/abs\/2304.08485"},{"key":"e_1_3_3_1_25_2","unstructured":"LLNL. [n. d.]. Variorum. Variorum. https:\/\/variorum.readthedocs.io\/"},{"key":"e_1_3_3_1_26_2","unstructured":"lmsys. 2023. LLaVA: Large Language and Vision Assistant. https:\/\/github.com\/haotian-liu\/LLaVA. Visual Instruct 150K Dataset."},{"key":"e_1_3_3_1_27_2","unstructured":"lmsys. 2023. Vicuna: An Open-Source Chatbot Impressing GPT-4. https:\/\/lmsys.org\/blog\/2023-03-30-vicuna\/. Accessed: 2025-12-15."},{"key":"e_1_3_3_1_28_2","doi-asserted-by":"publisher","unstructured":"Angel Melguizo Ra\u00fal Katz and Juan Jung. 2026. Can AI grow green? Evidence of a Kuznets curve among AI renewable energies and emissions. Energy Policy 208 (2026) 114883. 10.1016\/j.enpol.2025.114883","DOI":"10.1016\/j.enpol.2025.114883"},{"key":"e_1_3_3_1_29_2","doi-asserted-by":"publisher","DOI":"10.1145\/3295500.3356191"},{"key":"e_1_3_3_1_30_2","doi-asserted-by":"crossref","unstructured":"Ibrahim Niftiyev and et al.2025. The intersection of artificial intelligence (AI) green management and sustainability. E3S Web of Conferences 608 (2025) 05019. https:\/\/www.e3s-conferences.org\/articles\/e3sconf\/pdf\/2025\/08\/e3sconf_eenviro2024_05019.pdf Accessed: 2025-07-25.","DOI":"10.1051\/e3sconf\/202560805019"},{"key":"e_1_3_3_1_31_2","unstructured":"NVIDIA Corporation. 2024. NVIDIA Data Center GPU Manager (DCGM). NVIDIA Developer. https:\/\/developer.nvidia.com\/dcgm Accessed: 2025-07-25."},{"key":"e_1_3_3_1_32_2","doi-asserted-by":"publisher","unstructured":"Mihrimah Ozkan Lily Pompa Md\u00a0Shaihan Bin Iqbal Yiu Chan Daniel Morales Zixun Chen Handing Wang Lusha Gao and Sandra\u00a0Hernandez Gonzalez. 2025. Performance efficiency and cost analysis of wafer-scale AI accelerators vs. single-chip GPUs. Device 3 10 (2025) 100834. 10.1016\/j.device.2025.100834","DOI":"10.1016\/j.device.2025.100834"},{"key":"e_1_3_3_1_33_2","volume-title":"Proceedings of the 38th International Conference on Machine Learning","author":"Radford Alec","year":"2021","unstructured":"Alec Radford et\u00a0al. 2021. Learning Transferable Visual Models From Natural Language Supervision. In Proceedings of the 38th International Conference on Machine Learning. https:\/\/arxiv.org\/abs\/2103.00020"},{"key":"e_1_3_3_1_34_2","unstructured":"SmallCaps. 2025. IEA warns of surging energy needs as AI transforms global power landscape. SmallCaps (11 April 2025). https:\/\/smallcaps.com.au\/iea-surging-energy-needs-ai-global-power-landscape\/"},{"key":"e_1_3_3_1_35_2","unstructured":"Hugo Touvron and et al.2023. Llama 2: Open Foundation and Fine-Tuned Chat Models. arxiv:https:\/\/arXiv.org\/abs\/2307.09288\u00a0[cs.CL] https:\/\/arxiv.org\/abs\/2307.09288"},{"key":"e_1_3_3_1_36_2","unstructured":"Water Unite. 2025. Rising Water Demand from Data Centres: Addressing AI\u2019s Water Impact Through Innovation. Water Unite Blog (2025). https:\/\/www.waterunite.org\/blog\/post\/24500\/rising-water-demand-from-data-centres-addressing-ais-water-impact-through-innovation\/"}],"event":{"name":"SCA\/HPCAsiaWS 2026: SCA\/HPCAsia 2026 Workshops: Supercomputing Asia and International Conference on High Performance Computing in Asia Pacific Region Workshops","location":"Osaka , Japan","acronym":"SCA\/HPCAsiaWS 2026"},"container-title":["Proceedings of the Supercomputing Asia and International Conference on High Performance Computing in Asia Pacific Region Workshops"],"original-title":[],"deposited":{"date-parts":[[2026,1,22]],"date-time":"2026-01-22T13:38:35Z","timestamp":1769089115000},"score":1,"resource":{"primary":{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3784828.3785157"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,1,25]]},"references-count":35,"alternative-id":["10.1145\/3784828.3785157","10.1145\/3784828"],"URL":"https:\/\/doi.org\/10.1145\/3784828.3785157","relation":{},"subject":[],"published":{"date-parts":[[2026,1,25]]},"assertion":[{"value":"2026-01-25","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}