What's up with the data?

Jan H. Hansen
Jan H. Hansen
Joined: 7 Apr 05
Posts: 8
Credit: 668211
RAC: 0

First of all many thanks for

First of all many thanks for the update.

Ben>Weekly may be a bit much, since not that much changes from week to week.

I don't beleive this. You probably overestimate how significant changes are required to satisfy crunchers curiosity about progress.

An update need not be more than a up to date benchmark and estimated completion date. I assume this changes enough to ensure that you have something to report.

A weekly report like example below would go along way in preventing crunchers loosing interest in the project because of lack of information.

WEEKLY REPORT EXAMPLE
=====================

Since last update we have:
# more TFLOPS
# more CPU's
# more users

S3 is now 45% done with current performance projected S3 completion date is yyyy-MM-dd.

Roadmap:
yyyy-MM-dd - yyyy-MM-dd S3 crunching (sensitivity #)
yyyy-MM-dd - yyyy-MM-dd S3 analysis
yyyy-MM-dd - yyyy-MM-dd S4 crunching (sensitivity #)

rbpeake
rbpeake
Joined: 18 Jan 05
Posts: 266
Credit: 968000383
RAC: 1202191

First of all many thanks for

Message 11470 in response to message 11469

First of all many thanks for the update.

Ben>Weekly may be a bit much, since not that much changes from week to week.

I don't beleive this. You probably overestimate how significant changes are required to satisfy crunchers curiosity about progress.

An update need not be more than a up to date benchmark and estimated completion date. I assume this changes enough to ensure that you have something to report.

A weekly report like example below would go along way in preventing crunchers loosing interest in the project because of lack of information.

WEEKLY REPORT EXAMPLE
=====================

Since last update we have:
# more TFLOPS
# more CPU's
# more users

S3 is now 45% done with current performance projected S3 completion date is yyyy-MM-dd.

Roadmap:
yyyy-MM-dd - yyyy-MM-dd S3 crunching (sensitivity #)
yyyy-MM-dd - yyyy-MM-dd S3 analysis
yyyy-MM-dd - yyyy-MM-dd S4 crunching (sensitivity #)

This is a great idea!

We crunchers like to know we are making some progress with our work. Otherwise it is just work unit after work unit, day after day, week after week. We ask, are we getting somewhere, are we accomplishing anything other than just generating points?

We would know how the project is going, and where it is heading. We would definitely then be more likely to stick around for the ride, and not jump to another project. We would see our efforts are paying off, and directly advancing the project and scientific discovery in general.

So this weekly status update idea is wonderful!

jdhammer
jdhammer
Joined: 22 Jan 05
Posts: 4
Credit: 39124
RAC: 0

I think this is a great idea

I think this is a great idea as well. Minimal effort for huge reward in my opinion. I just hope that the Einstein@Home team (and all other D.C. project teams) don't take our crunching power for granted. I WILL contribute whether or not there's a weekly, monthly, or whatever status page, but it's nice to know that the people running the show are willing to do a little extra work to keep us in the loop.....I think I speak for almost anybody when I say even those few numbers plugged into a chart will keep us more than happy.

gravywavy
gravywavy
Joined: 22 Jan 05
Posts: 392
Credit: 68962
RAC: 0

A regular post like this one

Message 11472 in response to message 11460

A regular post like this one is probably one of the most important things that the developers could do to "reward" the faithful user community for their hard work and dedicated contribution to the cause. [...]

Anyway, thanks very much Ben. The information was very much appreciated.

yes, I'll add my support to this as well.

I'd also like to point out the use of the rating system.

At the time of writing the original posting in this thread already has a rating of +12 - thanks to those who added that feedback.

If you don't know how to applaud a posting, just click on the little plus icon at the foot of the posting: you will see the rating move up one.

If you liked the first post, and you also agree with Gary, you'd maybe want to uprate both the first posting and Gary's posting in this thread. Takes just two mouse clicks, m one for each posting, and is much quicker than composing a reply, and the equivalent of clapping or shouting 'hear hear' at a live meeting.

Some fair minded contributors also uprate a comment they disagree with because they think it raises an important question or is particularly well phrased. In which case they would want to add a written posting as well.

~~gravywavy

hoarfrost
hoarfrost
Joined: 9 Feb 05
Posts: 207
Credit: 92407747
RAC: 109045

WEEKLY REPORT

Message 11473 in response to message 11469


WEEKLY REPORT EXAMPLE
=====================

Since last update we have:
# more TFLOPS
# more CPU's
# more users

S3 is now 45% done with current performance projected S3 completion date is yyyy-MM-dd.

Roadmap:
yyyy-MM-dd - yyyy-MM-dd S3 crunching (sensitivity #)
yyyy-MM-dd - yyyy-MM-dd S3 analysis
yyyy-MM-dd - yyyy-MM-dd S4 crunching (sensitivity #)

Good idea!

I would be happy, if could read a weekly "progress reports" about state of data processing and "science reports" about scientific achievements (when science get an interesting results - not weekly, of course :)).

To progress reports, perhaps, good to add a graphic map of data processing - for example, vertical axis - numbers of data blocks (H1_XXXX.X) and on horizontal axis - returned results. Or many other graphic representations.

Ben Owen
Ben Owen
Joined: 21 Dec 04
Posts: 117
Credit: 65695060
RAC: 3121

Hi folks, It's good to

Hi folks,

It's good to know there's a "market" for more of a progress meter. Some of it has already been in place for a while.

If you haven't looked at the "Server Status" link on the main page, do so now. The most interesting number to the people on this thread is probably at the bottom of the middle column, titled "floating point speed in last 7 days". That's a measure of the real computing power you all have been putting in lately. Right now it's a factor of 3 down from the maximum number because not all the hosts are running. I don't know how this compares to other distributed computing projects, but I gather it's universal that some people sign up but don't run it or (like me) stop it on an older machine after a while.

Bruce is working on a much more graphic and interesting summary of the S3 analysis, including measures of scientific interest (and descriptions of how the science is actually done).

Stay tuned,
Ben

[AF>Amis de la mer] Eoerl
[AF>Amis de la ...
Joined: 5 Jun 05
Posts: 4
Credit: 5887
RAC: 0

Hi to begin with, thanks a

Hi
to begin with, thanks a lot for your reports ben, it helps to feel there's "human inside" ! I presume you don't have so much time to make us huge reports, or too much statistic comparisons, but reports of scientific or technical progress could do (or so i think)...
I had a question (this may not be the good place (?) ) : what distinguish clearly Seti and Einstein@home maths ? all in all, the spectrum used may not be so different (?) (Seti in Radioelectric waves, Einstein in gravitationnal waves -if we can compare those waves) even if the means to get those spectrums have nothing in common. So are the maths proceeded differently ? If they are not, could we imagine that both Einstein@home and Seti could have somewhat the same client, to proceed each kinf of waves the same way (and for Einstein@home to beneficiate of Seti's much longer experience). thanks, and sorry wether this question has no point

debugas
debugas
Joined: 11 Nov 04
Posts: 170
Credit: 77331
RAC: 0

In other words we are still

In other words we are still doing instruments callibration.

Jan H. Hansen
Jan H. Hansen
Joined: 7 Apr 05
Posts: 8
Credit: 668211
RAC: 0

It's good to know there's a

Message 11477 in response to message 11474

It's good to know there's a "market" for more of a progress meter. Some of it has already been in place for a while.

If you haven't looked at the "Server Status" link on the main page, do so now. The most interesting number to the people on this thread is probably at the bottom of the middle column, titled "floating point speed in last 7 days". That's a measure of the real computing power you all have been putting in lately. Right now it's a factor of 3 down from the maximum number because not all the hosts are running. I don't know how this compares to other distributed computing projects, but I gather it's universal that some people sign up but don't run it or (like me) stop it on an older machine after a while.

I didn't ask for a progress meter, I asked for a progress report. I appreciate you trying to understand what we are asking for, but I think you may be missing the point. If you re-read my post, that got fairly positive feedback, it should be pretty obvious that "floating point speed in last 7 days"-number isn't exactly what I was asking for.

You really need to improve the way you communicate with people who are donating their CPU cycles if you want to ensure their continued support.

Bruce is working on a much more graphic and interesting summary of the S3 analysis, including measures of scientific interest (and descriptions of how the science is actually done).

This sort of scientific article is very welcome indeed, but I suspect that the low frequency that you will be able to make these means you will need to take other steps to keep people interested.

Ben Owen
Ben Owen
Joined: 21 Dec 04
Posts: 117
Credit: 65695060
RAC: 3121

Hi folks, Again I'm trying

Hi folks,

Again I'm trying to make an omnibus answer to the thread.

Debugas:

E@H isn't really doing instrumental calibration. That's a pretty specific and limited thing, along the lines of "If there is a strain of X, possibly from a gravitational wave, this voltmeter will read Y at frequency Z." That doesn't burn a lot of cycles, although it does burn a lot of scientist-hours. Finding all the artifacts in the data is a different problem, which basically has to be done by analyzing the data - and that does burn cycles. The fake signals calibrate the efficiency of the analysis, but that is a complicated mix of mathematics and software and statistics that - again - can't be done without actually analyzing the data. If the instrument were the sort you get in fiction - 'scuse me, textbooks - with stationary Gaussian noise, that wouldn't be needed. But LIGO is a real life machine with much more complicated behavior than any book and so these extra steps are needed.

Eoerl(?):

I don't actually know how SETI@Home works. I suspect that their analysis is based on Fourier transforms, as is ours - or rather ours is lobotomized bits of several transforms stitched together. They probably have some things in common, but not enough to cut and paste from one client to another.

Jan:

I realize that the Server Status page isn't everything you were asking for, but it does have some of it. Like number of new users, though that is hardcoded to a one day interval. The numbers for total work units and work units completed are there too. From the numbers that are there you can work out percent completion and estimated completion time, though it would be nice to have the web page do it for you. I don't actually know how that page works, though I'll ask around.

The fundamental issue remains the fact that this is still a pretty new project with a chronic labor shortage. We'd like to have more progress reports rather than bug reports, but we're having to build it all as we go along.

I'm glad to talk about science issues on a more or less weekly basis, even though science results aren't going to be coming out that quickly. It's easier to answer people's specific questions than make up my own, so ask away.

Hope this helps,
Ben

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.