Chicken and egg
You need the GPUs to power fusion, but you also need the fusion to power the GPUs.
The UK Atomic Energy Authority (UKAEA) has recruited Intel and the University of Cambridge for the compute resources it needs to develop Britain's prototype nuclear fusion reactor – including building a "digital twin" of the design to help with testing. The Spherical Tokamak for Energy Production (STEP) is Britain's plan to …
"there just isn't enough time to build everything using a traditional iterative process of designing and then testing all the subsystems"
We've just had a striking example of what happens when you don't design then test all subsystems. I wonder what that forbodes for this ?
...Inside a top secret FUSION research team finance meeting:
Project Sponsor: Right team, we need to find a way of getting more money out of government, so we can keep up with our research jamboree, er I mean, scientific experiment
Project Manager: Good thinking. If we get this right, we could be securing our selves gainful employment until we retire
Chair: That's right. We're looking for at least another 25 years.
Worker #1: But won't they realise we've got nothing to show from the billions we've spent already?
PM: Apart from an enormously upscaled metal ringed doughnut and a very small lightening bolt.
PM: No chance. We;ll just tell them it's really complicated, use the word "quantum" a lot.
W#2: We did that last time.
PM: Did we?
W#1: Yeah, and we even threw in some virtual reality terminology
PM: Blimey. Hmm, OK any ideas?
W#3: How about Artificial Intelligence
PM: Not sure, could use it a bit, but we don't want politicians to think that Terminator will be in charge of the national grid one day
Chair: Good point. Also, if we do go AI, it might work out the problem for us, then we'll be out of jobs before retirement.
W#1: Why don't we talk about simulation models. Then we'll need an enormous spend to replicate everything we've done so far, but this time as a model inside a computer.
W#2: yeah, and then because we'll need thousands of GPUs, what with the global chip shortage, we will legitimately not be able to make any progress at all
PM: Genius! We'll get paid for achieving bugger all for years to come!
The problem is, fusion power is just another form of heat engine. To bring this debate up to date we need to go back to the beginning, where within my lifetime, 79 years, the majority of vehicles powered by either petrol or diesel were produced, and it is that aspect of this debate that has been, dare I say it, suppressed. For example, our weather related problems have nothing to do with beef production on grass; instead it is entirely caused by vehicles powered by heat engines, and exactly the same heat engine systems associated with the production of electricity; because all such energy is produced by what are known as “Heat Engines”. Burn a gallon of petrol, or diesel, produce electricity by nuclear power, all of the energy production is trammelled by the limitations of an inefficient heat to energy cycle of roughly 35%, (give or take a few percent depending upon the particular mechanism; piston engines in vehicles; steam turbines in power stations). The balance of the heat is immediately distributed back into the surrounding atmosphere. Mind you, not to forget that, all the energy created in any form of heat engine, always becomes heat eventually; the vehicle by movement, the electricity by distribution and the production of other forms of heat. All energy in the heat engines eventually becomes a heat input to the atmosphere.
Not wanting anyone to try burning a gallon of petrol in their garden, instead we need someone to set into motion a demonstration of, say, a six lane motorway full of vehicles each travelling at, say, 70 miles and hour, showing an illustration of the fuel being burnt, roughly 2 gallons per mile, above each vehicle. Now take 1,495 billion vehicles including trucks https://hedgescompany.com/blog/2021/06/how-many-cars-are-there-in-the-world/ and work it all out. THAT is the primary problem that it would seem no one wants to fully illustrate; and that does not include aircraft, then add the nuclear, (because we are being told that nuclear is the answer to global warming, when in fact it is just another form of heat engine, producing heat to create steam to power steam turbines; as also all the research into fusion power, is again, all about another form of heat engine). This time heating to some 10 million degrees to again turn water into steam to power turbines . . . same as; same as.
Humanity has no option but to abandon the use of heat engines to produce energy in whatever form; vehicle power, or electricity generation, if it wishes to survive. That is the real challenge facing all of us. Trying to raise funds for fusion power research is another example of a classic Red Herring specifically designed to avoid the debate about where the heat is coming from … heat Engines.
Surely you can put some numbers on this with back-of-the envelope calculations. You don't need to set up that experimental road, or even use a GPU. What is the per square km heat output of human activity on average? How does it compare with the per square km insolation from the Sun? Genuine question.
>” . The problem, he opined, is that a fusion reactor is an incredibly complex, strongly coupled system and the models for how fusion power plants operate are limited in their accuracy.”
Translation: We don’t really know what is going on in a fusion power plant to really understand what is going on and so model it accurately.
Hence expect someone has decided a sprinkling of AI pixie dust wil fill the holes in our.understanding and magically be able create more accurate simulations….
Being less cynical, as constraints-based programming and other deterministic programming techniques fall under the AI nomenclature, I would write a funding bid today sprinkled with the term “AI”, as it is clear politicians et al are prepared to throw money at “AI”.