Conference item icon

Conference item

Evaluating and comparing the potentials in primary response for GPU and CPU data centers

Abstract:

The rapid growth of Large Language Models (LLMs) and Artificial Intelligence (AI) has transformed traditional CPU-centric Data Centers (DaCe) into more powerdemanding GPU DaCes. Previous work has explored methods to reduce energy costs and carbon emissions in GPU DaCes. However, there remains a gap in understanding the potential of GPU DaCes for providing primary response, a crucial ancillary service for stabilizing the power system. Drawing on real-world job traces from a GPU-intensive DaCe operated by SenseTime and a CPU-intensive DaCe at Oak Ridge National Laboratory, we developed a mixed-integer linear programming model to assess the DaCe flexibility potentials considering individual jobs’ characteristics. We show that the GPU DaCe possesses a larger flexibility for delivering primary responses compared to the CPU DaCe. Furthermore, the GPU DaCe exhibits lower variability in flexibility across different times of the day and over a 7-month evaluation horizon, making them more dependable and stable sources for offering primary response.

Publication status:
Published
Peer review status:
Peer reviewed

Actions


Access Document


Files:
Publisher copy:
10.1109/pesgm51994.2024.10689061

Authors


More by this author
Institution:
University of Oxford
Division:
MPLS
Department:
Engineering Science
Role:
Author
ORCID:
0000-0002-5015-8661


More from this funder
Funder identifier:
https://ror.org/0439y7842
Grant:
EP/T028564/1
More from this funder
Funder identifier:
https://ror.org/019ya6433
Grant:
EP/S031901/1


Publisher:
IEEE
Host title:
2024 IEEE Power & Energy Society General Meeting (PESGM)
Journal:
2024 IEEE Power & Energy Society General Meeting (PESGM) More from this journal
Publication date:
2024-10-04
Acceptance date:
2024-07-21
Event title:
2024 IEEE Power & Energy Society General Meeting (PESGM)
Event location:
Seattle
Event website:
https://pes-gm.org/seattle-2024/
Event start date:
2024-07-21
Event end date:
2024-07-25
DOI:
EISSN:
1944-9933
ISSN:
1944-9925
EISBN:
9798350381832
ISBN:
9798350381849


Language:
English
Keywords:
Pubs id:
2037626
Local pid:
pubs:2037626
Deposit date:
2024-10-10

Terms of use



Views and Downloads






If you are the owner of this record, you can report an update to it here: Report update to this record

TO TOP