Read more.They calculated Pi to 62.8 trillion figures in 108 days and nine hours.
Read more.They calculated Pi to 62.8 trillion figures in 108 days and nine hours.
Why stop there? Can they not go to 70 or even 100 trillion figures?
How much did this all cost, though, and what ways will it benefit humanity over other things that could have been researched like... cost-effective planetary terraforming, or curing any number of diseases?
_______________________________________________________________________
Originally Posted by Mark Tyson
Read the damn article
Prof. Dr. Heiko Rölke, head of DAViS, said that preparing for the record attempt has been valuable in optimising computer processes making researchers aware of weak points in the infrastructure, and preparing them for future power-intensive data and computing tasks. Research partners will also benefit from the experience in their data analytics and simulations problems in the fields of RNA analysis, flow simulations and text analysis
I suspect the point is why didn't they put that optimisation effort towards something more useful if they were optimising infrastructure. There are any number of poorly funded, yet valuable scientific projects who would have donated their datasets for the effort and waited as long as it took for the results.
It's all about how many appropriate buzzwords you put in the grant,what bandwagon you can jump on,and who you know. However,as some of the best scientists by their very nature,are not the socially apt people,you can see where this is going(they can't sell themselves as well). If it happens to be good research.....!
The hardware used by DAViS is listed here.
In brief the computing system featured:
●2x AMD Epyc 7542 CPUs with 32 cores each
●1TB RAM
●510TB storage using a mix of HDD and SSD tech for swapping, Pi storage and OS
Well, to answer the question, the latest 5700G would take 109,648,800 seconds to calculate pi to 62.8 trillion digits.
That's about 1269(.083) days, or about three and a half years.
Do I get a prize?
If the Earth is a sphere how do you travel to the ends of it?
Whilst the research behind this is valuable, the exercise is not.
If they really want to see who is fastest, why not just have a competition to see who can get to 1 billion fastest / in the lowest energy use (say with a limit of 24 hours max run time)?
I think people may be overreacting to the news just a little bit. As posted by mtyson, this is the result from a single dual-socket system, not a whole supercomputer facility.
1 billion digits takes seconds with y-cruncher on even a desktop CPU. http://www.numberworld.org/y-crunche...charts/1b.html
Last edited by watercooled; 18-08-2021 at 03:47 PM.
All of these highest-precision-mathematical-constant records are the same - calculating the number is easy, storing it is not. 62.8 trillion decimal places is about 26TB
I already did what you have commanded, hence asking the damn question.
I know what they claim are the benefits... but those supposed benefits have only been described in very vague terms as a subjective assertion, rather than actually being quantified and - as per my question - What benefit does this bring compared to the other things I mentioned.
I don't want to know what they think, or what nominalisations their PR bod came up with - If that was all it took to satisfy me, I'd believe every word out of every MP's mouth!
I want actual values and a quantified position.
_______________________________________________________________________
Originally Posted by Mark Tyson
As I understand it, nobody knows for sure. Theory suggests the number is infinite, but it's sure beyond me knowing how to go about proving it until/unless someone finds the last number. Which they won't, if it is infinite.
What benefit? Why, the answer to the ultimate question of life, the universe and everything, of course.
Oh wait, that's been done, innit? 42, IIRC.
Beats me, then.
A lesson learned from PeterB about dignity in adversity, so Peter, In Memorium, "Onwards and Upwards".
There are currently 1 users browsing this thread. (0 members and 1 guests)