Fastest US Supercomputer Enlisted in Fight Against Coronavirus (bloomberg.com) 46
The fastest supercomputer in the U.S. is being put to work in the search for a vaccine to prevent the coronavirus and treat those infected by it. From a report: The Summit, housed in the U.S. Energy Department's Oak Ridge National Laboratory in Tennessee, is capable of 200,000 trillion calculations per second. It is being used to analyze health data as part of the Covid-19 Insights Partnership announced Tuesday by the agency as well as the departments of Veterans affairs and Health and Human Services. "Summit's unmatched capacity to analyze massive integrated datasets and divine insights will help researchers identify and advance potential treatments and enhance outcomes for Covid-19 patients with unprecedented speed," the agencies said in a statement. The Energy Department said earlier this year that its computers were being used to help the Centers for Disease Control and Prevention and the World Heath Organization conduct modeling on the virus.
Re: (Score:2, Insightful)
I wonder how much carbon this sucker belches into the atmosphere!
None apparently.. It uses the same electricity as your Electric Car...
I wish they would not do this (Score:2)
This is not the way to sell super computing. indeed it's like putting a computer in a superman costume and calling it "super". Big computers are best used for big problems not just large amounts of bad data with no useful predictive models.
it's embarrassing to people who do do big computing. Or really any computing at all
Re: (Score:2)
Actually, you can't tell from the summary. It *could* be valuable.
E.g. one thing that needs to be done is compare all COVID patients, tagged by severity of case, against the drugs that they're currently taking to see if any make a difference. Of course, some of the positive results will be noise, but it gives you places to look
Other people are doing molecular modeling, matching known drugs shape/electric charge/etc. against the shapes adopted by COVID pieces. That's another approach that might work.
There
Re: (Score:2)
Oak Ridge often uses experimental power sources on-site, and renewable power is available for purchase from the local power company, so maybe none?
And the Answer is: (Score:3)
Wear masks, social distance, no team sports, and wait for a vaccine.
Re: (Score:2)
Wear masks, social distance, no team sports, and wait for a vaccine.
I thought it was "The only way to win is not to play."
Re: (Score:1)
Show me the data from a randomized control trial and I'll believe it to the power to which it answers the study aims.
No data?--we call that a "guess".
Re: (Score:2)
Re: (Score:2)
You linked to an opinion piece, dipshit.
Re: (Score:2)
Re: (Score:1)
Re: (Score:2)
Re: (Score:1)
And also, 42.
Re: (Score:2)
Why? There are plenty of team sports where you are social distancing anyways. And there are a few of them in the Olympics too. Table Tennis (singles), regular tennis, curling, gymnastics, swimming (and with ceftain limitations, synchronized), diving, skiing, snowboarding, etc.
Just no team sports where everyone on the team is on the play area and where half the goal is to get close to one another.
I think motor racing sports were one of the first to be allowed on again, though I think the appeal
And will it help? No. (Score:3, Interesting)
Pretty much a stunt. May contribute some small things that could have done in other ways as well. If we knew enough to make vaccines a computational problem, we would have eradicated all infectious diseases a long time ago.
Re: (Score:3)
Pretty much a stunt. May contribute some small things that could have done in other ways as well.
Not a stunt at all. Molecular biology is now to the point that supercomputer molecular modelling can be used for a number of powerful techniques that couldn't be done any other way.
One of them is rapidly testing large numbers of candidate drug molecules against the molecular structure of virus components, to see which might bind to them in ways that would interfere with their normal operation.
Re: (Score:2)
I am aware of that. Since you have limited capabilities for the real-world tests that must come afterwards, this is pretty meaningless. You just get a larger back-log in an early step of the overall process.
Taking a break from Nuclear weapons production? (Score:2)
Ah, so they are using the super computers for COVID work now?
I guess that shows how important this is.. Giving up Nuclear Bomb simulation for this? interesting...
Re: (Score:1)
Re: (Score:1)
Re: (Score:2)
More telling is that they preferred to devote the last seven months to simulations of the maximally efficient ways to slaughter people, and only now has it occurred to any of them to use it to prevent some amount of deaths.
There are at least a dozen other supercomputers already involved in this research, this is nothing more than "Look at me too" for publicity.
Re: (Score:3)
I actually use this computer for research, and I can tell you that it is used to do tons of things that have nothing to do with "slaughtering people".
Look under "Scientific Impact" here: https://www.olcf.ornl.gov/summ... [ornl.gov] and, yes, those are real research project which take up huge amounts of time on the computer.
Also: this computer has been used for COVID since at least March - they were simulating other aspects of the virus to help in vaccine creation. This is just a new initiative using the computer.
For
Re: (Score:3)
Yeah, my bad. After I had read through the thread a little further I realized which machine this was and thought "Oops.".
folding@home (Score:1)
Re: (Score:1)
Re: (Score:2)
Re: (Score:2)
And if you want to throw CPU power at it, you can contribute to the OpenPandemics COVID19 project on World Community Grid. They now have work units for ARM and ARM64 CPUs too.
Re: folding@home (Score:2)
Hmn (Score:2, Interesting)
took awhile, didn't it?
Re: (Score:2, Insightful)
quite
Re: (Score:2)
No. Beginning in March it was used to help model the virus. Now it is being used in other ways.
Re: (Score:2)
Beginning in March it was used to help model the virus. Now it is being used in other ways.
I recall, about that time, a report that Lawrence Livermore's supercomputer facility's crunch had been loaned out for anti-covid work: to a project testing molecular target sites on the virus versus a library of the active ingredients of already-approved drugs, looking for existing drugs that might also work to suppress the virus.
So more than one nuclear research site's supercomputers are on the job.
Re: (Score:2)
So more than one nuclear research site's supercomputers are on the job.
MANY more than one lab. Shag found the announcement and posted a link to it just below. See "New, or follow-up to March?"
New, or follow-up to March? (Score:2)
Way back on March 23, the White House announced [whitehouse.gov] the "COVID-19 High Performance Computing Consortium [covid19-hp...ortium.org]" which included Summit [xsede.org] and a bunch of other supercomputers in government, industry and academia.
The announcement [va.gov] of this new "Covid-19 Insights Partnership" mentions the earlier consortium, but I can't tell whether this is just another piece of code being run as part of all that, or whether the consortium is losing Summit to the partnership.
I see (Score:3)
The US Administration has the biggest, fastest Computer of the world and it will save us all, it's beautiful, tremendous!
I wonder how... (Score:2)
How would this supercomputer compare against the swarm of folding@home?
Re: (Score:2)
The swarm became the most powerful supercomputer in the world by a wide margin when the pandemic started.
If we already have the vaccine... (Score:1)
Re: (Score:1)
Re: (Score:2)
Most vaccines that reach testing level 3 eventually fail. Prior to that most that reach testing level 2 fail before getting to level 3. To shortcut things many of the current vaccine endeavors are combining levels 2 and 3.
That said, vaccines are also of different quality. Some are more effective than others, some produce lasting immunity, others temporary immunity. Etc. Don't expect a good vaccine to be the first one developed, it doesn't usually happen that way. Usually the first vaccine that passes
Just be smart (Score:1)
Faster Computers (Score:2)
Faster Computers run bad models to achieve incorrect results faster than slow computers running the same bad models arrive at the same bad incorrect results.
Effort should be spent on useful (and perhaps correct) modelling, the calculations themselves are no more complicated than can be achieved with pencil and paper and a hand calculator of the sort commonly found in breakfast cereal boxes.
Being able to process garbage faster does not improve the quality of the garbage.
Patent Reform Suspended (Score:2)