PDA

View Full Version : Number of structures per batch?



mighty
05-31-2002, 11:26 AM
Hi!

On my Athlon XP 2100+ the program makes 3000 structures per batch. On my 500MHz pentium II (or III - I can never remember) it makes 5000 per batch.

What determines the number of structures per batch?

Regards
Ole

PS: Needless to say that both machines are running with the latest version of the program.

Richard Clyne
05-31-2002, 01:23 PM
Don't have an aswer to your question, but all my machines are still running at 5,000 structures per batch.

They are all using the latest software and use the no network switch.

FoBoT
05-31-2002, 02:44 PM
if your client is doing 5000 instead of 3000, I would totally delete the software and do a reinstall/re-unzip

its supposed to be 3000, so perhaps you have some mixed up files or something, or the update didn't complete correctly

Richard Clyne
05-31-2002, 02:57 PM
This was a new install. Prior to the new protein I erased my old client, downloaded the latest client, and once the new protein was available I ran the client. The update was completed and the rest is history.

I would like a comment from the peeps who designed the software to see if this is going to cause a problem.

I would be willing to try a fresh install on one machine, but not all five (56K DUN).

If there is going to be a problem with there being 5,000 instead of 3,000 structures per fileset then I will just switch back to seti.

Jodie
05-31-2002, 03:08 PM
ROFL! Hunting Aliens... Tooo funny! Maybe it's just me - but a project in three dimensional space rather than one in four dimensional space-time seems a bit more practical. Has anyone ever bothered to factor-in the temporal scales of speed-of-light limitation of radio communication in the hunt? Or the odds that a civilization would still exist that emitted a signal hundreds of millions of years ago? Or the odds that searching a miniscule fraction of the sky for a tremendously miniscule period of time would yield anything? Or the odds that the terribly limited frequency ranges searched would be utilized by a sufficiently developed civilization some Galaxies away? Or that the tremendous gravitational effects on radio waves might in fact cause significant bend?

Nothing but a PR gimmick for SETI to get some notice. A PR gimmick that got out-of-hand...

Ok, moving on -

I'm not running any of my clients with a -df switch - does changing the (effective) server timeout of the upload cause a larger periodicity betwixt uploads as well? Anyone running -df seeing 3k/upload attempt?

ulv
05-31-2002, 03:20 PM
I believe Howard said that 5000 would be the number if you're not on the net, otherwise you'll get the numbers from DF (in the Technical section).

I run all boxes with -df, they are always online, they upload after 3000.

Richard Clyne
05-31-2002, 03:26 PM
Okay, Jodie I take it hell will have to freeze over before you would consider crunching a few seti units.

Each to their own, thats what I say.

If someone feels searching for little green men is making more appropriate use of their time and money then that is up to them.

I have never slagged anyone for what ever project they wish to participate in, no matter how I feel about it.

I don't expect to be slagged off for the projects I paricipate in.

Jodie
05-31-2002, 03:28 PM
I didn't attack you. I attacked SETI. Unless you officially represent the SETI Group, don't take it quite so personally.

And I think your mystery was effectively solved between the end of my post and Ulv's, right?

Jodie
05-31-2002, 03:31 PM
Heck, let's go one step further - I have a firewall monitor (an HP Pavilion zt1130 1.13G laptop) that is on DUN/private. It routes to the 'Net over a point-to-point dialup.

I'll go ahead and DL the client now (should be done with the download in, oh, a week ;) ), disconnect, and start it up and let you know what I see...

[edit: I had it on here already - so I'm just doing an 'update' - hopefully it's for practical purposes the same]

Sami
05-31-2002, 03:32 PM
At least none of my boxes were able to download a new protein but went on processing the old one indefinitely, periodically trying to contact the server. Eventually I got frustrated and downloaded the new program version from the web site, deleted everything old and installed it again to all the machines. It worked, but I'm worried that other people may have the same problem.

Richard Clyne
05-31-2002, 03:44 PM
Originally posted by Jodie
I didn't attack you. I attacked SETI. Unless you officially represent the SETI Group, don't take it quite so personally.

And I think your mystery was effectively solved between the end of my post and Ulv's, right?

Okay thank you for clearing up that misunderstanding - The project is a waste of time, but the people taking part are not wasting their time.:confused: :rolleyes:

So until ths problem is resolved I am away to waste my time searching for little green men.:rolleyes:

Digital Parasite
05-31-2002, 03:57 PM
Mighty: Just in case you got lost/confused with all the SETI talk, online clients will do 3000 structures per upload on this protein and offline clients will do 5000.

Jodie: It doesn't matter if you use -df or not, the "block" size is still 3000 for online and 5000 for offline.

Jeff.

Richard Clyne
05-31-2002, 04:02 PM
Originally posted by Digital Parasite
Mighty: Just in case you got lost/confused with all the SETI talk, online clients will do 3000 structures per upload on this protein and offline clients will do 5000.

Jodie: It doesn't matter if you use -df or not, the "block" size is still 3000 for online and 5000 for offline.

Jeff.

Cheers Jeff for your reply - so I don't need to go and re-install all my clients as 5,000 offline is expected :D

The little green men will have to wait for another day :D

Jodie
05-31-2002, 04:04 PM
If you like, you can stop searching for little green men now. ;)

I think we have an experimental answer:

The lab rats:

P3-Mobile 750 Sony Vaio/64Meg/Internal 56k modem.
P3-Celeron Mobile 1.13G HP Pavilion/256Meg/Internal 56k modem.

The test: Two different phone lines, one connecting to Earthlink/Sprint, the other on private dial-up routed to the 'Net. One a fresh download, one an update. Download and install. Kickoff the program whilst connected to the 'Net. Observed next update in 3000 units for each. Hit 'Q' and let it upload. Added -df switch to each. Started them up. Observed 3000 units to next upload for each. Hit 'Q' and let them upload.

Disconnected from the 'Net on each. Kicked off the program.

It attempted to search for new versions and tried to spawn a new connection, which I denied. It then used current handle and displayed 5000 to next upload. On both laptops.

Conclusion:

I believe we can safely conclude that if you are currently connected to the 'Net - you update in 3000 structures. If you aren't connected (ascertained by the software by trying to search for new version) - then you attempt upload in 5000 structures.

Make sense?

Richard Clyne
05-31-2002, 04:32 PM
Jodie, thank you for taking the time to run your tests. I am glad to see their is some harmony in the results.

I would still be interested to hear from the peeps in charge of developing the software, just to confirm that the findings are correct. 3,000 structures per fileset for online and 5,000 structures per fileset for us peeps having to work offline.

I do not wish to fold to my hearts content only to find out my results are useless (would have been better spending that time looking for little green men:rolleyes: ).

I would also be interested to know what to expect when you go from working offline to online - will the number of structures drop from 5,000 to 3,000 per upload or will it stay at 5,000?

ulv
05-31-2002, 04:51 PM
Richard: As I said before, Howard answered this question in the Technical Support section, you can trust him;)

Richard Clyne
06-01-2002, 05:34 AM
Originally posted by ulv
Richard: As I said before, Howard answered this question in the Technical Support section, you can trust him;)

Please accept my apolgies. It was late at night and I something else on my mind, and to be honest I missed your post:o

:cheers: