Hello you fine Internet folks,
Today's article is on the memory subsystem of Arrow Lake from the perspective of Lion Cove and looking at the behavior of Lion Cove compared to Skymont and Zen 5.
Hope y'all enjoy!
open.substack.com/pub/chipsand...
old.chipsandcheese.com/2025/01/05/a...
Couldn't find a Starter Pack for green computing, so gave it a go! 🌱💻 No doubt it's missing many great ones... please add! Hoping to include anyone discussing environmental impacts of compute (not just pure CompSci), the broader the perspectives the better! #greencomputing
go.bsky.app/CCE1mJe
It's been great so far!
In Glasgow for #LOCO2024
An incredibly exciting programme to look forward to 🌱
Our Low Carbon Computing workshop #LOCO2024 is taking place now in Glasgow and online! We already had a nice keynote chat with Anne Currie and now we are in the first regular session on Energy Efficiency! 💻🔬🌱
Full room at our first hybrid workshop on Low Carbon Computing #LOCO2024 at @uofglasgow.bsky.social, hopefully the first of many gatherings of a community working together for low-carbon and sustainable computing. Registration is still open if you want to join online: www.sicsa.ac.uk/loco/loco2024/
The University cloisters covered in twinkling lights.
It’s feeling festive in the cloisters! ✨
#TeamUofG
The problem with talking about utilization is that the concept can be tweaked to sound good for whatever audience needs the metric. For example, someone might say 💯% utilization is a node being used. But look closer? The app might only be hitting 40% of the GPU. Or the power ratio might be terrible.