Dr. Peter Garraghan presented his work on sustainable Machine Learning Systems at the British Computing Society (BCS) Real Artificial Intelligence group.
Guest Lecture at BT
Dr. Peter Garraghan presented his research on sustainable systems at the BT Applied Research Thought Leadership event.
Presented at UK Systems Research Challenges Workshop
Dr. Peter Garraghan presented his work titled
The Efficiency Death-March: The Unintended Consequences of Large-scale Systems Research upon Climate Change at the Fifth Annual UK Systems Research Challenges Workshop.
Guest Lecture at The University of Bristol
Dr. Peter Garraghan presented a guest lecture on Experiment Instrumentation to the EPSRC Doctoral Training Centre on Security at The University of Bristol.
Accepted IEEE TDSC paper on Cloud Security
Our paper titled Cross-VM Network Channel Attacks and Countermeasures within Cloud Computing Environments has been accepted for publication within IEEE Transactions on Dependable and Secure Computing.
Awarded EPSRC Early-career Fellowship
Peter Garraghan has been awarded a prestigious EPSRC Early-career Fellowship titled Reducing the Global ICT Footprint via Self-adaptive Large-scale ICT Systems.
This £1m project aims to radically redefine the notion of ICT sustainability, and transform how we engineer and use digital infrastructure in the face of environmental change.
Interference-aware DL scheduling paper presented at ICA3PP
PhD student Gingfung Yeung presented his paper titled Horus: An Interference-Aware Resource Manager for Deep Learning Systems, at ICA3PP, New York.
New machine learning systems arrived
Our new Nvdia Tesla V100 GPUs machines have now been setup within the EDS datacenter laboratory, and will be used for researching machine learning system security and energy usage.
(Note: They are surprisingly loud)
First year PhD panel passed
Congratulations to PhD student Petter Terenius for passing his first year appraisal panel.
Papers presented at USENIX HotCloud’20
PhD students Gingfung Yeung and James Bulman presented their work at HotCloud’20 co-located at USENIX ATC 2020.
Towards GPU Utilization Prediction for Cloud Deep Learning
A Cloud Gaming Framework for Dynamic Graphical Rendering Towards Achieving Distributed Game Engines