Posted 10 days ago by
Mike Hewson
Good question. I think if we consider the whole range of computational tasks that are associated with our current GW search, we currently already have the part that is best suited for volunteer computing out in the field. The main factor is the ratio of computational cost in terms of CPU hours to the volume of data that needs to get in and out of the computation, and the amount of latency that we can afford to get back the result (is it blocking the next stage?). So (currently) I cannot think of a better way to make use of the volunteers' resources, but that's just me of course ;-).
Thanks HB, I thought it would be something like that : otherwise it would have probably been already done.
... so I'm not really sure how difficult it would be to port (say) BRP4 to this platform. Yes, it supports OpenCL, but that doesn't mean it's as easy as recompiling....
Yes one can support any language, the question is whether that will yield some efficiency advantage using that hardware. My initial goal - with the board yet to be seen or touched, mind you - is to get some FFT going with, say, a CUDA-FFT library style interface ( setup/plan/pre-calculate then execute ). Learn from there ....
Cheers, Mike.
Posted 11 days ago by
Mike Hewson
So far so good with presenting serial characters to the OGLFT interface, on a continuous overnite run! I'll leave it running for another three/four days before I claim victory. :-) :-)
I suppose to be exact it may not be the OGLFT code per se, but maybe the FreeType2 which it calls, or for that matter MinGW which does the cross compile ( not trouble whatsoever on Linux ). And of course Windows itself. Any one - or interaction(s) thereof - could be responsible. All I have done is to note that it sometimes falls over within Face::draw() calls where the argument is const char*. I guess assumptions can readily date when upgrades occur. The most recent OGLFT version ( 0.9 ) was produced in 2003 ....
Now in the process of tracking this I've also removed much OpenGL display list usage, replacing with more forward compatible constructs ( display lists are not supported, i.e. illegal, in v3.2+ contexts ).
Cheers, Mike.
Posted 11 days ago by
Mike Hewson
Update : obscure bug being tracked down. Heap overrun within OGLFT library code in Windows builds only, and alas then only sometimes. Related to use of C-style string constructs ie. char * and an assumed '\0' ending for the Face::draw() function. It took me five (5) weeks to define just that. Workaround : attempting to present single characters serially to the OGLFT interface. Sigh. I think there is some 80/20 rule or somesuch : 20% of the bugs take 80% of the time. I'd make it 98/02 ..... :-)
@Robert : I too have done some online classes, which probably gave me the impression I could succeed in this venture. Programming is far more than merely typing code in some language paradigm though! There's a level of rigor required and good tools are essential. My sympathies to our E@H developers ....
Cheers, Mike.
Posted 14 days ago by
Mike Hewson
I am wondering if there is some indication of whatever might be wrong in the tasks' stdout / stderr outputs. Something that I could compare for a given WU between my task that is found invalid and another that is found valid.
Excellent idea, in fact that's one reason why such logs exist. I'll have a look when I get home. :-)
Cheers, Mike.
Posted 16 days ago by
Mike Hewson
Cool. My name is on the same page as Stephen Hawking. That'd be a first for him I expect. :-) :-0
Cheers, Mike.
Posted 16 days ago by
Mike Hewson
.... that information which is encoded in my body structure would! :-) :-)
Well, that information thus merged with all other captured stuff .....
Hereabouts
The metre. Our current basic scientific unit of length in the MKS system ( Metre/Kilogram/Seconds ). Some use CGS ( Centimeters/Grams/Seconds ). Originally defined as the length of a pendulum with a given period, then one ten millionth of an Earth quadrant i.e. pole to equator - though never measured exactly as such - then a series of metal bars of particular production and situation. Now defined as that distance traveled by light in a certain time*, though in practice remains a certain count of wavelengths of specific emitted standard light sources. Whatever. Take it as read that we know what a metre is and can when needed line other things up against that to make statements like "the distance from Melbourne to Frankfurt is ... " accurate/sensible/consistent.
So we can hence forth measure any and all things here on Earth - the word 'geometry' means exactly that - and even these days have very accurate measurements by remote calibration. This is otherwise known as the Global Positioning System which relies upon highly specified coordinate entries, base points if you like, combined with time of flight measurement of light signals in the form of radio waves. Think of the speaking clock ( an 'old' service but still operating ) per phone emitting the likes of 'at the third stroke it will be ten forty five am and fifty seconds .... beep .... beep .... beep'. Suppose I heard that very same message but by several distinct paths, and the 'origin' of each path ( satellite in this case ) encoding it's position when the time signal was emitted, then you have ( with suitable General Relativistic corrections for the variation in the Earth's gravitational field hereabouts ) sufficient information to deduce/calculate where you are. For instance I could hear that third beep earlier from one satellite than from another. Of course in practice it isn't quite like that but that's the simplest explanation I can think of without going on about spacetime events and reference frames et al. Naturally there is all manner of issues as regards accuracy for any given measurement instance eg. how many distinct satellite contributed signals were enjoined.
An obvious question here is : how does the satellite 'know' where it is? It cannot inform me with my GPS unit as to my whereabouts if it doesn't itself have a well specified position from moment to moment. You can sense the inevitable answer : it must check with base camp(s). They do and indeed crosscheck against each others' positions too. Again complex in the detail.
The deeper point here is that we are at the first rungs of a ladder. As one steps to another scale one still refers back to - calibrates against - a prior scale. You can do this in either direction of course ie. go up to higher scales ( longer lengths ) or to lower scales ( smaller lengths ). In fact it is not only distance that is subject to that ie. ditto in principle for other quantities. We are human and thus to make sense of our investigations, for our brains, the information has to return to our scale - which notably is of the order of a few metres, a few kilograms and a few seconds - to be absorbed and pondered upon**. So the hard drive which is storing this post is going to be doing that according to some agreed pro-forma/encoding, but we are not personally inspecting the small magnetic domains on the platter that serve to represent such information. We have constructed circuitry to do that, and said circuitry was manufactured using other calibrated instruments to enable that. Et cetera ... in a chain of mutually calibrated procedures and devices.
Einstein was exceptional in that he really took to task the explanation/practice of measurement. This is what is frequently found annoying by newcomers to relativity discussions. Why do they go on about rulers and clocks and observers ? Isn't that obvious ? What's the issue here ? Well you can't actually consistently discuss space without mentioning time, you'll hit it sooner or later, so we ought not go onwards to cosmic scales without tackling :
The Speed Of Light
Cheers, Mike.
*Interesting that we've come full circle to a time based definition of length, which of course regresses to what is such a time interval anyway? Note that for small amplitudes all pendulums of the same length will have the same period, this being first ( well recorded/apocryphal ) noted by Galileo sitting in church watching the candelabras suspended from the ceiling. It won't depend on the mass attached to the end of the pendulum but does depend on the local gravity field ie. acceleration due to gravity at the point of operation of the pendulum. Interestingly the current length definition has a caveat regarding general relativity essentially acknowledging that yes, clocks in different parts of a gravitational field will run at different rates.
**Alas some navel gazers have taken this to support their contention that science is 'socially defined' and cannot make absolute comments about anything. See my earlier comments about Craig Lowndes, but apart from that it is simply the case that scientifically slanted slugs, or questioning quokkas, or intelligent motivated asteroids are all going to fail to reach the speed of light if they try to travel that way. My money will be on the photons arriving first. Science may have various social/language encodings - somewhat like the choice of ASCII for computer character data - but an independent universe ( and the same one in each case ) does exist regardless. You can always regress to measurement. A black hole will swallow up all comers. Of course I appreciate that how people interact and behave in common endeavours - like scientific studies - has lots of cultural spins and group thinks. But let's not confuse that with falling to the ground when you trip.
( edit ) "You can always regress to measurement". This probably is the line in the sand with respect other disciplines of study that claim the 'science' label. Bringing back information to human scale does not lock us to a human centric view. In fact technology has allowed many surrogate viewpoints to arise eg. an image of the Earth from the Moon, or writing little messages with an electron microscope. You can think outside your own head.
Posted 17 days ago by
Mike Hewson
... that all the 'info/matter/whatever' is stored at the entrance around the opening and nothingness is what is inside.
The basic idea here is that quantum mechanics comes into play and allows for the black hole to radiate particles from the horizon, hence a mass/energy loss. Eventually the hole will 'evaporate' at a rate which is inversely proportional to mass. So tiny holes go pop with a quick flash of radiation, but larger ones ever so slowly. Applying thermodynamic rules one can view the hole as lying within the heat bath of the cosmic microwave background ( CMB ) radiation. It's a cool bath actually. In any case the hole has an effective temperature going like the inverse of it's mass ie. big is cool, small is hot. Weird huh ? Currently the numbers are such that stellar mass plus black holes are cooler than the CMB, so it'll be a while before the CMB glow fades such that they will be warmer than their surroundings. Then they'll lose energy and begin to dissipate by an excess of radiated loss over absorbed gain.
Now the information/entropy content of the hole is not going to be visible to the rest of the universe until it begins significantly emitting. Hawking's interpretation is thus that the info is stored until a way later time : some humungous figure like 10 to the sixtieth power years ( we're only around 10 to the power of nine years currently ). Could be more. Long time.
You could heuristically explain this as saying that since, from the perspective of a long distance from the hole, the horizon has 'frozen' features - time dilation is so severe that clocks ( thus frequencies too ) go to a zero rate. So when the horizon is shrunk with hole evaporation, so nearby spacetime returns to flat, horizon clocks come back up to speed and the information 'wakes up' and returns to the remainder of the universe. So in a weird sense a black hole operates like a time machine : you get to go to a very distant future time with only a short elapsed proper time. Well I wouldn't personally survive intact but that information which is encoded in my body structure would! :-) :-)
Cheers, Mike.
Posted 22 days ago by
Mike Hewson
Exactly. BOINC essentially parallelises, and the server side validates ( currently at E@H using specific validation per returned result, but also via quorums ). BOINC is the framework for distribution but doesn't determine content. I'm just musing about whether more work - out of the entire data processing pipeline - can be shifted to volunteer clients, and whether that is useful/acceptable/pragmatic etc. A basic BOINC design 'rule' is that only server-side work can be 'trusted', so with our current workflow there are already mechanisms in place to determine if client work can be trusted ( in the scientific/cognitive sense ).
I'm just shooting the breeze here. The thought came up when I was thinking of parallel computing in the specific with Adapteva ie. how to apply that ( if at all ) to E@H. I do fully anticipate developer wrath here .... :-)
Cheers, Mike.
Posted 22 days ago by
Mike Hewson
Interesting news from this project:
http://www.hpcwire.com/hpcwire/2013-04-23/adapteva_shows_off_$99_supercomputer_boards.html
Sigh .... I've been waiting at the gate for the postman every day. :-)
I've been thinking - obviously in generalities at this stage - of how to parallelise ( further ) the GW work. Maybe as the WU's currently stand then there's probably not much scope - maybe a bit quicker FFT's say. But then I thought that we have already parallelised, in that E@H by it's distributed nature is effectively doing that already. Soooo ....
[vaulting ambition overleaping itself]
To what extent can work currently labelled as 'pre-' and 'post-' ( with respect to distributed WU allocation ) be shifted from server side to the volunteer milieu? Could there be scope for creating new classes of work units that do that stuff ? Or is there little to no advantage, accounting for outlay to setup and maintain, compared to status quo ? Is it the type of work that needs a level of validation beyond our present criteria for volunteer submitted results? Or on the 'post-' side is it the case that the results go many different ways for further ( shall I say 'ad-hoc' ) analysis? Or is it none of my business ? Will the implementing developers groan collectively and then group-fund a hitman to shut me up? ;-) :=)
[/vaulting ambition overleaping itself]
Cheers, Mike.
Posted 23 days ago by
Mike Hewson
Hi Gary ! :-)
One idea that immediately springs to mind for yourself - though not necessarily solving any problems for others - is to use a proxy server ( ie. caching type intercepting outbound requests ). For your herd you could dedicate one rig for that role : using Apache HTTP Server and/or Apache FTP Server say, suitably configured and of course setting the proxy dialog appropriately within BOINC preferences. How and when you trim the cache is up to you.
Cheers, Mike.