UCY-CompSci
@UCompsci
High-performance computing with a mission: contribute to a better world.
You might like
The George Gilder Technology Report : “The End of the Microchip” gilderreport.com/the-end-of-the… My take: “The Beginning of Locality”. The center of gravity is shifting toward integration and locality. We are seeing wafer‑scale engines, box‑scale UMA, and rack‑scale fabrics that cut…
Walking towards the future… 🚀 & the HPC-AI Synergies Lab awaits 😉 — join us to learn how you can use MLX on Apple Silicon for your AI/ML applications in Engineering & beyond. Nov. 5th is the official inauguration of the new Engineering Complex at @UCYOfficial 🎉 @awnihannun…
Huge thanks @Korata_hiu for kicking the tires on Kourkoutas‑β! 🦎 Love hearing it works on SDXL/T2I and becomes your default vs fixed β₂. Paper → arxiv.org/abs/2508.12996 Code → github.com/sck-at-ucy/kbe… PyPI → pypi.org/project/kbeta P.S. Availability diagnostics & efficiency…
Awesome work! I tested it on T2I models (e.g., SDXL) and implemented it in my library of advanced optimizers (I had to lag the EMA of the gradient norm by one step to make it compatible with fused backpass). It yielded very good results, it will be my default instead of fixed β₂
Happy to have contributed to the development/validation of this product. Fun memories from the collaboration. Solid engineering team on the @aptar side.
Gentle mist. Targeted delivery. Proven science. Aptar Pharma's #PureHale® helps parents treat toddlers naturally - no chemicals, no fuss. ✨ Reaches mouth, nose, throat & pharynx ✨ Ideal for colds, allergies & dry nose Read #AptarPharma's article for more ow.ly/NjsV50X2H6n
🚀 🎉Just dropped our first public release of Kourkoutas-β! An Adam-style optimiser with dynamic β₂ memory for bursty gradients. 🔗 Paper: arxiv.org/abs/2508.12996 🔗 Code: github.com/sck-at-ucy/kbe… 🔗 PyPI: pypi.org/project/kbeta 🦎🌞 “we code in MLX” @awnihannun @angeloskath
CDA Mangis: Pleased to join the Stanford Alumni Association Cyprus (SAAC) dinner. Thanks to cofounders Suzi Abdel-Malak & Effie Kokkinou. SAAC promotes networking among Stanford alumni in Cyprus, strengthening the 🇺🇸🇨🇾 connection to the Stanford University family.
🥳🚀 1st installment of Mac Studios M4 Max for the “HPC+AI Synergies” teaching lab has arrived. 🚀🚀 Benefits: 👉🏻Ultra low energy footprint 👉🏻 Quiet student-friendly environment for hands on experimentation w/ parallel connectivity topologies 👉🏻Students get trained on machines…
I’m sooo 🦎🦎happy 🦎🦎w/ Kourkoutas-β. Completed 1st full test on training a heavy Transformer model as PDE surrogate in data-driven mode. Kourkoutas-β blows vanilla Adam out of the desert, especially for smaller training datasets & quantized models where the grads tend to be…
The Deep Dive 🚀🚀 podcast on our preprint of “Beyond Language: Applying MLX Transformers to Engineering Physics” doi.org/10.48550/arXiv… Code available at github.com/sck-at-ucy/MLX… @awnihannun @angeloskath #AI #ArtificialIntelligence #ML #MLX #NeuralNetworks
The second Podcast 🎧 in our series dives into our efforts @UCompsci with collaborators to create a computationally efficient model of the deep lung 🫁 by taking some clever shortcuts: The cited article is doi.org/10.1016/j.jaer… #AerosolScience #InhalationTherapies #LungHealth
Starting today we initiate a series of science podcasts 🎧 on X showcasing some of the most exciting research output of @UCompsci @respihub & collaborators. Thanks to Google for making this possible. These podcasts are themselves a testimony of the real value of AI 🚀🚀
🫁🌊 Collaborating with Josué Sznitman on this Annual Review was fun & rewarding. It is finally live in early release @AnnualReviews : Multiscale Modeling of Respiratory Transport Phenomena and Intersubject Variability | Annual Reviews - go.shr.lc/3Z5KXKj
Thermo Therapist with CoolProp is being released in the GPT store for my Engineering Thermo students. It explains concepts, uses a custom API to access CoolProp, solves problems & will convince you that you should ❤️ Thermo. Not responsible for any Transference issues 😂 @OpenAI
🚀🚀🚀🐢 We verified that if you mix an M1 Max w/ several M2 Ultra you can almost recover full performance of using all M2U by proportionally reducing batch size on M1, increasing it on M2, keeping total # of epochs the same. Might depend on case/size. @awnihannun @angeloskath
Xgrid lives — a new project resurrects the promise of Apple's dead clustering software @KassinosS @awnihannun @angeloskath appleinsider.com/articles/24/06…
appleinsider.com
AppleInsider.com
Evoking the old Xgrid days, a new project connects Mac Studios together with Thunderbolt cables, and uses them in tandem for massively parallel computing tasks.
Xgrid lives -- a new project resurrects the promise of Apple's dead clustering software appleinsider.com/articles/24/06… #Apple
appleinsider.com
AppleInsider.com
Evoking the old Xgrid days, a new project connects Mac Studios together with Thunderbolt cables, and uses them in tandem for massively parallel computing tasks.
Some like CalDigit Thunderbolt 4 Element Hub do offer 40 Gb/s on all TB ports. But as correctly pointed out by @elegyals you can connect up to 7 MacStudio M2 Ultra using either an iPad as monitor over WiFi or other non TB monitor (over usb or hdmi)
ahh I see, but should it be fast on a daisy chain loop which uses 2 USB ports each, with arbitrary size? given that training will most likely be ring AllReduce. I doubt Hub can be full throttled to 40G/(s*port), they are rarely designed for it
In the lab we saw a trend of improving speedup ratio as we went from 1 to 2 to 3 to 4 nodes, I’m suspecting it has to do with more efficient use of lazy evaluation in each node. To be fair we should average performance over many more tries for each setup & compare averages
United States Trends
- 1. Justin Fields 9,188 posts
- 2. Jets 65.9K posts
- 3. Patriots 141K posts
- 4. Drake Maye 18.5K posts
- 5. Henderson 20.9K posts
- 6. Jalen Johnson 4,867 posts
- 7. AD Mitchell 2,273 posts
- 8. Judge 178K posts
- 9. Pats 13.9K posts
- 10. #FinallyOverIt N/A
- 11. Diggs 9,914 posts
- 12. Cal Raleigh 7,297 posts
- 13. #criticalrolespoilers 2,169 posts
- 14. #TNFonPrime 3,244 posts
- 15. Santana 13.4K posts
- 16. #GreysAnatomy 2,042 posts
- 17. Mike Vrabel 5,768 posts
- 18. Purdue 8,883 posts
- 19. #TNAiMPACT 5,711 posts
- 20. Metchie 1,312 posts
You might like
Something went wrong.
Something went wrong.