Arecibo graphic question

I'll take a stab at an explanation of how I think this works. If someone more knowledgeable finds error please do enlighten me! =)

I seem to remember an explanation somewhere on these message boards of the "Power spectrum" to be a real time representation of the signals/noise in the data as it's being processed. That is, higher bars would indicate some type of signal, but before this can be said to be a signal from a pulsar the data has to be post-processed back at the project together with data from multiple WUs.

Edit: Further down the page here in the Cafe I found this post from Mike about the Power spectrum.

Now over to the stderr output.
I believe stderr stands for "Standard error" and the output is a log of messages from the application that can be used to diagnose errors.

I've taken a part of a log from one of the WUs that you have processed and if we break it down I would say that this is what we get:

The part of the log above contains information about the Boinc version and the resource used to process this WU, in this case a "GeForce GTX 550 Ti".

Then follows some information about the data which is to be processed and at the end of this block is the search parameters used.

Above is some more memory information after the GPU setup is complete.

Then follows information about the actual processing of the WU and we can see that it checkpoints one time before completing.

Then the whole thing repeats seven times as one WU consists of 8 individual tasks.
If something where to go wrong during processing then hopefully this log will contain information about this to help with diagnosing the error.

The actual results of the data processing is stored in a separate file and uploaded to the project after the WU is completed and the contents of that file is not shown in any log.

Hi!

I'm very thankful for your methodical and simple to understand explanation.

As I thought myself (but was far from being sure about it) the 'role' of the POWER SPECTRUM is the 'real time' graphic representation of signal/noise ratio.

So if you are correct (and my thoughts on this are the same) I can draw only one conclusion out of this:

If an WU intermittently shows high activity in the Power spectrum then it is somehow more 'promising' in finding something as opposed if there was no activity at all.

That is a conclusion based merely on pure logic and data I have avaliable (almost none). So ANYBODY with more knowledge, PLEASE correct me, if this assumption is wrong.

In the case of the stderr output I'm really grateful to you. Most of the things you explained excellent and methodical I already knew (not all the details), but the part I DID NOT KNOW, helped me VERY much and saved me a lot of time (as I was searching for something, that did not exist at all :) namely:

So again:

Thank you very much for your kind, quick response and explanation.

Denis

P.S.
Any other comments on the subject, written with very simplified words below, are welcomed.

Is there any way for the cruncher to know if a particular WU has higher chances of finding something significant than other?

P.P.S.
As I was writing, I did not notice, that beside Holmis I got also a reply from Mike Hewson. So thank you very much to BOTH of you.

Thank you very much for your response!

But it is not only that, but something MUCH MORE important!

As a physician I'm very much aware how hard it is sometimes to explain things to students or even lay-people, who are for example interested about a particular disease, a physiological process or something similar.

It happens quite often, that some things, you are asked about and which seem in your head as being 'common sense' and as a result of such mentality, your answer (as precise it might be to your fellow colleagues) contains too much complicated information, with the end-result, that the person, who was asking, is in the end:

1. 'Overflown' with too much conflicting data, which he does not understand;
2. More confused as he was before the question was even asked.

And I'm not even talking about 'rocket' science here, nor am I an 'academic' in my area. But nevertheless I MUST never forget, that I studied 7 years, to gain BASIC knowledge (because if my 'vocabulary' would not contain the sentence: 'I don't know.' sometimes, I'd be lying to myself and doing more harm than good).

I can only imagine, how someone with your knowledge, possesses much patience and pedagogic abilities (because in this scientific area is even a larger gap between lay-man and professionals, namely - said simplified - mathematics) to explain VERY COMPLEX things, to a wide 'public', with varying and diverse levels of education.

That is the MAIN reason, that I'm grateful for your time. It is hard enough to be a good physicist (or mathematician), but nothing in comparison with the skill of being able to explain difficult topics in a way, that the person, who is asking, knows at the 'end of the day' more (and becomes maybe even more curious), opposed to receiving lots of data, which (at least for him) is useless.

For me, that is the MOST SIGNIFICANT difference between an average and excellent teacher.

Regards.

Denis

P.S.
I excuse myself for being so long, as English is not my primary language and I'm sometimes almost struggling to find the right words.

It might be simpler to think of 1.42GHz as 'photons of a certain energy' ( or within some defined band ) with the variation in time of the flux/number of those being received as our primary signal data. There is the famous connection between photon frequency ( as disclosed by experiments that demonstrate wave effects like diffraction and interference ) and photon energy ( as exhibited by the photo-electric effect or atomic electron transitions ) which Albert Einstein championed ( although presaged by Max Planck's treatment of radiation in cavities ) :

Energy = h * frequency

h = Planck's constant

More photons -> stronger signal, less photons -> weaker signal. The frequency mentioned in our work units is how rhythmic is the variation in that photon flux, not the choice of energy of the photons we choose to count. So presumably the ~1.42GHz is a technical choice related to how one designs, builds and operates an Aricebo dish.

You see, back at the pulsar there is this huge ( double ended ) photon 'hose' which pours out a massive amount of radiation at a wide range of energies. These are very unhealthy neighbourhoods to be around. Think microwave oven. Because of the spinning of the ( neutron ) star the hose is whipped around in a circle, and if the geometry is right with regards to it's pointing in 3D space combined with where we are then we get splashed intermittently by the photons from said hose. There are no doubt many of these objects that we will never be aware of as there are configurations aplenty where we never get wet. Some pulsars are labelled as 'intermittent' which is I think generally interpreted as the hose changing the plane of it's rotation ( due to the influence of nearby objects ) now and then. There are even 'binary pulsars', fascinating beasts, that have a pair of these guys ( each with their own hoses ) circling around each other and maybe sometimes blocking from us the other partner's signal .....

Think of your favorite lawn sprinkler ( or the usual lighthouse analogy ). So it's the frequency of the rotation of the sprinkler that we are seeking ie. how fast is this neutron star spinning? [ I'm skipping over a heap of arcane stuff I barely understand ... ]

I'm always a bit boggled by these objects : mass about that of the Sun ( give or take ), around 20km in diameter, a stupendous magnetic field, and some have been found to spin faster than a Ferrari's crankshaft at full throttle! :-)

Well if you see something sticking out a mile, away from the left/low frequency end, then don't let me crush your optimism! :-)

Cheers, Mike.