Will the United States have the world's fastest supercomputer in June 2022?
8
104Ṁ285
resolved Jun 1
Resolved
YES
Technology
Get
Ṁ1,000
to start trading!

🏅 Top traders

#NameTotal profit
1Ṁ52
2Ṁ20
3Ṁ11
4Ṁ8
5Ṁ7
Sort by:
Top500 June list came out, and Frontier is first: https://www.top500.org/lists/top500/2022/06/ resolve?
Haven't researched it myself, but Lawrence C. makes a compelling argument for 'maybe yes'
Happy to buy to ~55%. The question is basically about whether or not the new, currently under construction Frontier supercomputer will submit an HPL result to the Top500 list by May. - The question is resolved by checking the Top500 list, which updates twice a year in June and November with submission deadlines of May and October respectively. - The current #1 Top 500 list has the Fugaku supercomputer at 442 petaflops, while the proposed Frontier system will have somewhere north of 1 exaflop (it's not clear how to scale peak flops to flops on the Linpack benchmark used by Top500). - While China seems to have 2 exascale supercomputers already up for almost a year and a third to be completed soon, 1) their reported Flops are (probably?) less than the proposed Frontier system's and more importantly 2) they having not submitted any HPL results to Top500 in for either submission deadline in 2021, despite these systems being used on quantum computing results released in November 2021. This makes me think that the chance a non-US party submits a result faster than the Frontier system is quite small. (Call it <15%.) That being said, I'm pretty uncertain about if Frontier will be ready by then. If we look at how long the Summit supercomputer took to go from delivery to results, it began delivery in Late July/August 2017 and took #1 place on the June 2018 list (though it wouldn't reach full performance until mid August 2018 and its current top500 numbers until the November 2018 list). The Frontier supercomputer began delivery in August 2018, and it doesn't need to reach full performance to win the first place. I'd say there's a 2/3 chance they manage to submit an HPL result by this point, taking into account 1) the difficulty of the system and 2) the slightly tighter timeline than the Summit system. Multiplying together the numbers, we get ~55%.

People are also trading

Will the largest machine learning training run (in FLOP) as of the end of 2025 be in the United States?
86% chance
Will the US place restrictions on the total compute capacity individuals or companies are allowed to have before 2026?
2% chance
Will any supercomputer reach 1 zettaFLOP before 2035?
82% chance
From which countries will the 5 most performant quantum computer corporates in 2028 have originated?
Will someone from sundai.club get a Nvidia's Project Digits desktop AI supercomputer by EOY2025?
69% chance
Will the US Government implement "Universal Basic Compute" by 2030?
11% chance
Will the US implement a policy for rapid shutdown of large compute clusters and training runs by 2028?
33% chance
Will the largest machine learning training run (in FLOP) as of the end of 2040 be in the United States?
46% chance
Will the largest machine learning training run (in FLOP) as of the end of 2035 be in the United States?
69% chance
Will the largest machine learning training run (in FLOP) as of the end of 2030 be in the United States?
74% chance
© Manifold Markets, Inc.•Terms•Privacy