Examining which research is the most important?
30 Jan 2021 15:56:19 UTC
Topic 224661
(moderation:
I am starting this thread to get into the question of research priorities vs. credits awarded.
I am aware this topic is likely to "gore somebody's Ox". If you are one of the gored ones PLEASE restrain your temper and argue courteously!
In another thread, our very well informed Moderator offers the opinion that Gravity Wave research is the "gold" and Gamma-Ray research is the "silver" of the importance of this research.
If that is the axiom then why should Gamma-Ray research get a quantitative reward that is greater than the Gravity Wave research reward?
If Gravity Wave research is the most important of the two then why not re-arrange the rewards to encourage it?
I understand there is "history" and a reluctance to change the rewards but the rewards are "arbitrary" even if they have a specific rationale for how they were decided on. The rewards allow for short-term tracking productivity. To allow the volunteers to compete.
So if Gravity Wave research is the most important why not change the rewards so it will more likely to get the most computer resources?
Respectfully,
Tom M (moderately techy IT guy)
A Proud member of the O.F.A. (Old Farts Association). Be well, do good work, and keep in touch.® (Garrison Keillor)
Tom M wrote:In another
)
I'm only reiterating what the Director of this project says. You can read the full thing by looking at "About" under the Science tab. Here is a direct quote.
The LIGO detectors go through periodic upgrades. There can be times (years even) where there is no data to analyse. The Pulsar searches provide a backup for such times and there have been quite a number of previously unknown pulsars discovered as a result. Fast spinning, nearby pulsars are likely to provide the first detections of continuous GW, if the supposition that pulsars might display 'millimeter mountains' has any validity. The 'out-of-balance' mass, spinning at high speed should radiate continuous GW.
Pulsars in tight binary systems orbiting each other are also a target of interest. Finding such targets should make it computationally much easier to do a 'directed' search rather than an 'all sky' search. So the various pulsar searches are the perfect complement to the primary search.
The project seems quite happy to accept the results from whichever search a volunteer chooses to run. You have the freedom to choose whichever search interests you. Credits are just for fun if you choose to pay any attention to them. Anyone serious about measuring and optimising performance looks at average elapsed times for tasks known to have reproducible performance. Sometimes that's easy and sometimes that's hard to assess. Some of us make an effort to try to work out the various factors involved.
Cheers,
Gary.
Tom M wrote: I am starting
)
Another idea could be they have more of the Gamma-Ray tasks and need to get thru them and giving more credit for them does that for some people.
There are Boinc Projects paying huge credits for a small amount of time spent crunching, they have their reasons for it but don't always express them publicly. My personal guess is some want to make as big an impact as they can while they can. Other projects pay far fewer credits per task and their stated idea is to be a steady eddie and not have the peaks of a ton of people coming for the credits and then leaving. Giving out more credits can also be a bad thing for a project if they don't understand and prepare for the 'credit whores' who crunch just because they look at the numbers climb. Some projects have crashed, and some continue to crash, as more and more people bring more and more resources to the project, they just can't handle the load. That often leads to less credits being given out to slow down the number of new people signing up or those staying or the project going away because they just don't have the cash flow to keep buying newer and faster hardware.
There are also BRP tasks available here at Einstein for people with ARM(raspberry pi) devices, those tasks take over 10k seconds and they give out 62 credits for each one. Does that mean they don't need those tasks crunched at all, or barely? IMHO as long as Einstein doesn't go to the Seti plan of "credit new" then I'm okay with however they do it.
Gary Roberts wrote:The LIGO
)
[Aside]What would be really cool ( I dream on .... ) is a new pulsar detected by both methods at E@H, in either order. That would really cross validate alot of physics & methodology. Surely I can't be the first person to have thought of that.[/Aside]
Anyway, as for the idea of credit adjustment to favor GW searches over EM searches .... the OP makes a good argument. A counter though : credits per se are not the only possible goal. Maybe some contributors are going for a detection* by their own personal computer ( it seems, at least in the near term, that an EM detection is more likely than a GW ). That goal on a personal level is independent of relative credit assignments per se, even if the project does manipulate that subset of crunchers for which credit is the prime motivator.
Maybe, just maybe, the pulsars won't give us a GW detection ie. our searches keep pushing the upper bounds on emission ever downwards without a specific finding. Given that the history of physics contains some spectacular "null" results**, that could rewrite some basic theory ( or at least inform upon the equation of state for neutron stars ).
Cheers, Mike.
* But I think most of the low hanging fruit is long gone.
** Ever seen a lone quark ? Or measured the 'ether wind' ? Or measured the difference b/w gravitational and inertial mass ? Mind you, nullity can be a relative term eg. you never really 'see' a neutrino ( there are no neutrino tracks in detectors ) as their existence is only inferred by the behaviour of other particles, and in order to balance the books on some major conservation principles.
I have made this letter longer than usual because I lack the time to make it shorter ...
... and my other CPU is a Ryzen 5950X :-) Blaise Pascal
Mike Hewson wrote:** Ever
)
That sounds like a great "handle" shades of the Lone Ranger.
Tom M
A Proud member of the O.F.A. (Old Farts Association). Be well, do good work, and keep in touch.® (Garrison Keillor)
On the trail of the lonesome
)
On the trail of the lonesome quark.
Richard
[long winded aside, it's a
)
[long winded aside, it's a Friday with time to spare]
The origin of the use of the word quark is said to come from the phrase "Three quarks for muster Mark !", mentioned in James Joyce's Finnegan's Wake. Joyce doesn't define what a quark is. The three is significant though, that being the number within a proton or neutron. Richard Feynman had a related idea of partons, a useful way of looking at collisions involving hadrons eg. electrons hitting protons. It was Murray Gel-Mann who came up with the initial quark model. The symmetries in that model exclude single quarks and indeed none have been seen. It's a relativistic thingy. If you put enough energy into a collision in order separate a free quark from some bound state with others : you just wind up creating quark/anti-quark pairs which then can be detected in composite particles of two's and three's or as quark jets.
I am amazed that particle physicists can say so much about these wee little things via bashing them together and seeing what emerges. The everyday analogy would be to crash cars together at high velocity and deduce their design & structure by examining still photographs of all the pieces as they fly off.
The apex of this style of investigation is of course the discovery of the Higgs Boson. You don't see Higgs tracks in detectors either, one deduces their existence via statistical analysis of collisions of sufficiently high energy. The Higgs decays in many ways to more ordinary particles in a really brief instant. The analogy here is to deduce a grenade's pre-explosion structure by the shrapnel it produces.
One big hole in physics at the moment is how/where to place to place gravity in such particle schemes. One would think that just inventing a graviton - the supposed quantum of the gravitational field - would do it. But there is a plague of problems with that. For instance it turns out that one can't consistently localise energy in gravitational interactions, the energy is a property of the entire field, and that is implicit in General Relativity. Roughly speaking one can say "the Earth falls towards the Moon" or "the Moon falls towards the Earth" equally validly, so an observer on one of those bodies is going to measure the other body as the one doing the accelerating. This is supposedly true right down to the domain of small dimensions. Furthermore gravity is a distortion of the space & time measurements that different observers compare. The Standard Model of particle fields leave space and time alone with only Special Relativity corrections when comparing viewpoints.
As GR has been such a successful theory many are loathe to go up against it. My personal idea, just shooting the breeze here, is that there needs to be a theory of gravity intermediate b/w the cosmic and the atomic scales. There would be an explicit scaling factor that causes gravity to go to zero strength over small distances ( leaving the success the Standard model un-opposed ), but going to the standard inverse square approximation of Newton as separations increase ( leaving the success of GR un-opposed ). But the key thing is to construct a quantitative model, and that means some fancy mathematics. I have a name at least : the Duck Soup Model, because I duck the hard questions and leave a messy mathematical soup ...
[/long winded aside, it's a Friday with time to spare]
Cheers, Mike.
I have made this letter longer than usual because I lack the time to make it shorter ...
... and my other CPU is a Ryzen 5950X :-) Blaise Pascal
Duck Soup Model, What
)
Duck Soup Model,
What would Groucho Marx say to that?
Richard
That would have to be about
)
That would have to be about the silliest movie I've ever seen, but delightful at that. A witty political parody, the brothers at their apex. :-)
[on topic]Another reason to care about credit is for the sheer sake of optimising it. A challenge to push one's hardware and tinker for fun.[/on topic]
Cheers, Mike.
I have made this letter longer than usual because I lack the time to make it shorter ...
... and my other CPU is a Ryzen 5950X :-) Blaise Pascal
If it walks like a
)
If it walks like a quark.
Quacks like a quark.
And lays interactions like a quark.
Then it is Duck Soup?
And Thank you, Mike. I haven't been doing the Readers Digest reading on the tribbles with the GUT (Grand Unified Theory) and/or Standard Model. So I appreciate the commentary.
Tom M
A Proud member of the O.F.A. (Old Farts Association). Be well, do good work, and keep in touch.® (Garrison Keillor)
I am in the process of
)
I am in the process of transitioning an AMD 3950x CPU (16t/32c) to "all" GW CPU tasks (except for the gpus).
I have a gpu server that is running mostly GW gpu tasks (5 GPUs currently, 1 gpu offline).
I am hoping for some of that "gold" :)
Tom M
A Proud member of the O.F.A. (Old Farts Association). Be well, do good work, and keep in touch.® (Garrison Keillor)