"Report on the Fujitsu Fugaku (富岳) System", Jack Dongarra, Jun 22, 2020
https://www.dropbox.com/s/aqntdb43p6so0z5/fugaku-report.pdf?...
6,976 karma · joined September 25, 2014
Consulting: https://www.blopeur.com Intellectual Property meet AI : https://www.patentblitz.com
Blog: https://www.reflectionsofthevoid.com/ Medium: https://medium.com/@benoit.hudzia
Linkedin: https://www.linkedin.com/in/benoithudzia/
"Report on the Fujitsu Fugaku (富岳) System", Jack Dongarra, Jun 22, 2020
https://www.dropbox.com/s/aqntdb43p6so0z5/fugaku-report.pdf?...
This delivers a huge boost in bandwith.
HMB stand for high memory bandwidth. It offers up to 900 GB/s.
Now if you add the tofu interconnect on top you have a systems finely tuned for maximising data movement.
Remember : compute is cheap, communication is expensive.
You can have load of gpu and processors but if you can't feed them data fast enough they are useless.
"Microsoft has deemed C++ no longer acceptable for writing mission-critical software"
"70% of the CVEs originating at Microsoft are memory safety issues"
Yea.. so basically you have a catalogue but it's going to be up to the consumers or providers to find a way to stitch up a clearinghouse solution for dynamically orchestrated services from a multitude of providers. This is not only hard. It's near impossible. Without this crucial element, the platform is stillborn.
Why? Because it's damn near impossible to deliver such feature for dynamically orchestrated services from a federation of providers.
Sovereignity, interoperability, compliance is a nice sales pitch. But it seems that when it comes to getting paid or paying you are on your own. Imagine trying to pay for hundreds of services from companies spread across dozens of countries...
You have to remember that the HPC market is $35B today. HPE makes $3B a year alone from that, maybe more with Cray acquisition.
So it's no surprise that NVIDIA wants to position themselves on that market.
Plus, you have too look at the long game with MLX acquisition ( heavy player in the HPC market) and Cumulus. I wouldn't be surprised to see NVIDIA trying to bypass Intel/AMD completely and offer Direct to Interconnect device. Rather than the hybrid CPU + GPU box.
Starter Code Purdue-P Coprocessor : https://nanohub.org/resources/33337/
"...this brings into question the results of other cutting edge papers not covered in our experiments. It also raises doubts about the value of the hand-wavy theoretical explanations in metric learning papers."
> AUTHORS
> Nick SOHNEMANN, Future Candy
> Lasse Michael UFFRECHT, Future Candy
> Marie Constanze HARTKOPF, Future Candy
> Jette Paulina KRUSE, Future Candy
> Lena Marie de NOELLEN, Future Candy
Also WTF the EU is thinking to task what looks like a "house of lies" consulting agency for a Study Digital services
look at this presentation they did for EU .. I was expecting a Better of Ted jaberwooky at any moment in the slide deck .. -> https://www.europarl.europa.eu/cmsdata/196436/IMCO%20KEYNOTE...
"Overview of hls4ml project", Oct 16, 2019 https://indico.cern.ch/event/855626/contributions/3600304/at...
"Status & Future developments", Apr 22, 2020 https://indico.cern.ch/event/911672/contributions/3834385/at...
Github : https://github.com/hls-fpga-machine-learning/hls4ml