Comments on: Sky-High Hurdles, Clouded Judgements for IaaS at Exascale https://www.nextplatform.com/2023/09/27/sky-high-hurdles-clouded-judgements-for-iaas-at-exascale/ In-depth coverage of high-end computing at large enterprises, supercomputing centers, hyperscale data centers, and public clouds. Thu, 28 Sep 2023 13:07:14 +0000 hourly 1 https://wordpress.org/?v=6.7.1 By: HuMo https://www.nextplatform.com/2023/09/27/sky-high-hurdles-clouded-judgements-for-iaas-at-exascale/#comment-214222 Thu, 28 Sep 2023 00:57:03 +0000 https://www.nextplatform.com/?p=143047#comment-214222 Great summary and analysis of ORNL’s snorkel into the sky-high depths of airborne cloudy HPC gastronomy! One may have hoped that composing disaggregated heterogeneous cumulonimbuses into a single distributed computational unit could have provided us with a capability class leadership supercomputer of the zettaflopping variety … but, unfortunately, the survey says: not so fast Buster (for all the reasons given in this excellent TNP piece)!

It should be some time then before HPC sorcery brews sufficiently potent IaaS concoctions to jump the hurdles posed by the opaque fog of cloudy hardware nodes and proprietary network witches, as needed to summon-up the impending rise of ZettaCthulhu.

In the meantime though, DeLorean-styled qubit tech, associated with the fluxonium capacitor (FTF sandwich with transmon as dielectric), may well provide us with time-traveling computations, that output the correct answer, before the code is even run. Quite the opposite to Jeopardy-inspired LLM tech, where one has to first give the answer to the machine (or many, many, answers), and then try to guess, or prompt-engineer, a question that will make it re-produce that answer as its output! q^8

]]> By: Hubert https://www.nextplatform.com/2023/09/27/sky-high-hurdles-clouded-judgements-for-iaas-at-exascale/#comment-214218 Wed, 27 Sep 2023 22:57:17 +0000 https://www.nextplatform.com/?p=143047#comment-214218 Hah! The words “remarkable”, “intricate”, “amazing”, “tantalizing”, “astonishing”, and “vital”, occur nowhere (not once) in this 29-page ORNL report … I have to guess that its target audience consists of intelligent professionals (nice!). (eh-eh-eh!)

]]>