We have a nice RTS queue just in time for our Tuesday maintenance !! I'm so glad with all the improvements and upgrades that the system has had to make it so stable and able to handle so much more.

Thanks for the graph Kiska. I think this is a great place to talk about the how the system is doing when it isn't a panic situation.

Yep things look good.

---

What "feature" requests do you want me to implement?

If you could move the Validation values from the present graph (sahv8 Workunits) to the one above (sahv8 Results in Progress), as they are of similar magnitude- in the millions.
Where as the Assimilation, Deletion & Result_deletion values should all be zero (or very close to it) under normal conditions. Even when things go bad, they're usually in the thousands (occasionally the 100's of thousands when things go really, really bad). But never in the millions, as the Validation numbers always are.
Thanks for your efforts.

---

Grant
Darwin NT

Does anyone have an automated method for providing an average processing time on cpus for the wildly varying named tasks?

I am seeing under 40 minutes and slightly over and hour on the cpu side of my biggest box. I would be interested in seeing a report on that. But I don't really want it to be a manual process.

Tom

---

A proud member of the OFA (Old Farts Association).

Does anyone have an automated method for providing an average processing time on cpus for the wildly varying named tasks?

Remember that you have a file called 'job_log_setiathome.berkeley.edu.txt' in your BOINC data directory.

That contains a single line of information for every task you've ever run, since the machine was first attached to the project. Name, runtime, CPU time are all accessible.

The major fly in the ointment is that there's no data about which device a task was run on: you can only really perform the sort of analysis you're interested in on the log from a machine which runs SETI on one device only: CPU only, or a single GPU only. But it's a start.

Also remember that for Arecibo tasks, the name doesn't contain any information about the task complexity.

You can just scrape the info from the website. I do/have done it plenty. Then sort and analyze the info offline with excel or whatever tools you prefer.

You’re limited by whatever tasks still remain on the website though.

---

Seti@Home classic workunits: 29,492 CPU time: 134,419 hours

Hi Richard,

Remember that you have a file called 'job_log_setiathome.berkeley.edu.txt' in your BOINC data directory.

That contains a single line of information for every task you've ever run, since the machine was first attached to the project. Name, runtime, CPU time are all accessible.

WOW!!! Who's idea was it to do such a thing? I just loaded that file into a text editor and it took about half a minute or more to load. I though it crashed the PC, at first. My file is currently almost 14MB HUGE! That's just over 3 months worth. I'd hate to see that file after 3 years or even longer. You'd never get it loaded into a text editor! If I delete the file, (I really have no purpose for it), will BOINC recreate it and continue on? Do we have any kind of control over that file?

Have a great day! :)

Siran

---

CAPT Siran d'Vel'nahr - L L & P _\\//
Winders 11 OS? "What a piece of junk!" - L. Skywalker
"Logic is the cement of our civilization with which we ascend from chaos using reason as our guide." - T'Plana-hath

You can just scrape the info from the website. I do/have done it plenty. Then sort and analyze the info offline with excel or whatever tools you prefer.

You’re limited by whatever tasks still remain on the website though.

Hi Ian,

What does a website scraper look like? I have scrapers out in my Snap-On roll-away tool box in the garage, but I doubt any of them would work. ;)

Have a great day! :)

Siran

---

CAPT Siran d'Vel'nahr - L L & P _\\//
Winders 11 OS? "What a piece of junk!" - L. Skywalker
"Logic is the cement of our civilization with which we ascend from chaos using reason as our guide." - T'Plana-hath

Scraping is a common term for using a script to retrieve data from the web.

It doesn’t “look” like anything. It’s code.

---

Seti@Home classic workunits: 29,492 CPU time: 134,419 hours

My file is currently almost 14MB HUGE!

And mine's 50 MB - 427,309 lines.

I recommend Notepad++ - 'opens' (displays the first page of) the file instantly. Also copes with the *nix convention line endings - that's a lot of the reason for the slow opening with other Windows tools.

I strongly recommend using the file rather than web scraping - we need to protect our rather delicate SETI servers.

I've done enough fairly large scrapes (1000+ tasks), to say that it has little impact on the site/server.

either the server intentionally slows down page requests, or my code just isn't written in a way to accelerate the process. each task is on a new page, which forces you to constantly be loading new pages, which slows it all down to about 1 task/second.

---

Seti@Home classic workunits: 29,492 CPU time: 134,419 hours

That's still one database query per second. BOINC tries to aggregate multiple database accesses when possible - delaying reporting tasks until a batch can be process together, for example.

One can probably deduce CPU vs GPU for anything other then low power GPUs by simply looking at the run time. Now to work out what the format of the file is (unless I can find the key somewhere)

---

Bob Smith
Member of Seti PIPPS (Pluto is a Planet Protest Society)
Somewhere in the (un)known Universe?

My 'SETI data distribution history' is driven from those files via an Access (==SQL) database. I'll get you the table schema.

Bah - it's only queries where you can edit the SQL directly. Is this enough?

Ta.

So far worked out that 9th column is the task name (pretty obvious really!~) and 11th is the run-time (not so obvious).
For most folks that would be a good starter for ten (even when there are over a million entries in the list as one of mine has!)
I'm going to guess (so that Richard can put me right) that the first column is some sort of modified Julian date format, probably the run date & time - this is just based on the way it increments....

And Richard & I crossed paths mid flight - as usual

---

Bob Smith
Member of Seti PIPPS (Pluto is a Planet Protest Society)
Somewhere in the (un)known Universe?

Looking at the data those make sense.

Now to do some playing and see what happens.

---

Bob Smith
Member of Seti PIPPS (Pluto is a Planet Protest Society)
Somewhere in the (un)known Universe?

This is how I spec'd it

Remember that you have a file called 'job_log_setiathome.berkeley.edu.txt' in your BOINC data directory.
That contains a single line of information for every task you've ever run, since the machine was first attached to the project. Name, runtime, CPU time are all accessible.

. . Thanks for the additional information ...

Stephen

:)

My file is currently almost 14MB HUGE!

And mine's 50 MB - 427,309 lines.

I recommend Notepad++ - 'opens' (displays the first page of) the file instantly. Also copes with the *nix convention line endings - that's a lot of the reason for the slow opening with other Windows tools.

I strongly recommend using the file rather than web scraping - we need to protect our rather delicate SETI servers.

Hi Richard,

I'm not running BOINC on Winders, I'm running on Linux. I use the Notepad that is installed with Wine. Perhaps I'll try something other than Notepad. ;)

I have no clue on how to scrape the web/database/servers.

Have a great day! :)

Siran

---

CAPT Siran d'Vel'nahr - L L & P _\\//
Winders 11 OS? "What a piece of junk!" - L. Skywalker
"Logic is the cement of our civilization with which we ascend from chaos using reason as our guide." - T'Plana-hath

After 6 months, the file on my Linux machine is 37 MB, 274,673 lines long. It opened instantly in the native Ubuntu text editor, but did take several seconds to scroll down to the end.

I hesitate to try and open my 280MB file with text editor. What the heck . . . . opens instantly but takes two minutes to crawl the loading process bar to the end.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)