Between 21:10 and 21:50 approx GMT+2 I noticed some sudden change in the frequency block widths being displayed on the client screen, up to over an inch wide in a more uniform fashion than I've ever seen.

Example;


The varations in frequency block width lasted for 1-3 displays, before reverting back to normal width.
All this occured within approximately 1-5 minute intervals.
The widest blocks were displayed at 21:36 and 21:37 GMT+2.

I can provide full detailed information, as well as more detailed pictures and current sky position if necessary.

If this isn't something completely usual, I'd love to get info how to forward this, if it has any relevance at all.

Thanks in advance.

-Fili

---

The app runs at several different bandwidths or dispersion levels. Some levels can not be processed in some tasks. So it is probably nothing really unusual, just you have not noticed it before.

---

BOINC WIKI

BOINCing since 2002/12/8

Thanks for the reply.

You're probably right, as processing power needs would ^x at an alarming rate if all frequency bandwidths were to be maintained at a constant rate all the time.

I noticed the bandwiths were particularly wide when a new data chunk underwent processing. (at 0.0->0.1%).

Cheers.

---

... I noticed the bandwiths were particularly wide when a new data chunk underwent processing. (at 0.0->0.1%).

I'm not sure what you mean by that...

The s@h client runs through a whole series of searches for pulses and gaussians and all at various drift rates depending on the angle range (AR) that the telescope swept across the sky whilst that WU was being recorded. The different mix of tests for a WU for a particular AR greatly affects how long a host computer takes to run through that WU. Hence some of the interest that the credits fanatics have for what "AR" a WU has.

Note that seti is all a 'best guess' at what might be found and what might be distinguishable as 'artificial' to indicate "ET".

Meanwhile, this is a fantastic project in itself involving very many people with various discoveries waiting to be found aswel as ET.

Keep searchin',
Martin

---

See new freedom: Mageia Linux
Take a look for yourself: Linux Format
The Future is what We all make IT (GPLv3)

... I noticed the bandwiths were particularly wide when a new data chunk underwent processing. (at 0.0->0.1%).

I'm not sure what you mean by that...

Sorry for the large, and 2d-like display of the image, but here's an example taken from the midst of a chunk even, and not at the very start as I mentioned before.

The bands of each frequency are very wide, opposed to more normal diplays, where you can see as many as 500-700 individual frequency bands (guesstimate). This is what gained my initial interest in the phenomenon.

---

... I noticed the bandwiths were particularly wide when a new data chunk underwent processing. (at 0.0->0.1%).

I'm not sure what you mean by that...

Sorry for the large, and 2d-like display of the image, but here's an example taken from the midst of a chunk...

Good clear example, thanks.

The bands of each frequency are very wide, opposed to more normal diplays, where you can see as many as 500-700 individual frequency bands (guesstimate). This is what gained my initial interest in the phenomenon.

Analysis screenshot

Oh that's easy... And completely normal. Usually, that step is soo quick that I guess most people won't notice...

Look at the frequency resolution for that part of the search. You're on the very fast 1200Hz step. The analysis also runs through much finer steps which is what you're more familiar with seeing.

See:

What is the Screensaver doing NOW?

How SETI@Home Works

Fast Fourier Transform


Good question there! Hope that helps.

Happy crunchin',
Martin

(Not had the "screensaver" on for so long that almost forgot what it looks like!)

---

See new freedom: Mageia Linux
Take a look for yourself: Linux Format
The Future is what We all make IT (GPLv3)