Interesting seti BOINC stats: BBR & TA.

[TenaciousT]

Rank Name Members Recent average credit Total credit Country
3 BBR 42 1301.02 7595.10 None
16 TA 33 211.10 1424.00 None

They can't be allowed to be ahead of us in TWO different versions of the client!

I say we give em Seti2 but we hold onto our Seti1 position. My single BOINC machine is switching back...

Top BOINC teams

 

[TofBnT]

Sounds good to me.

If we can whomp them in 1 then we can rally together and whomp them in 2

 

[SNC]

Seti 2 I think will be a flop. I am glad to hear your switching the one back.

 

[networkman]

Originally posted by: SNC
Seti 2 I think will be a flop. I am glad to hear your switching the one back.

Roger that.

 

[Hellburner]

All the beta numbers will be zeroed out once S@H2 actually starts.

 

[UlricT]

Originally posted by: Hellburner
All the beta numbers will be zeroed out once S@H2 actually starts.

are you sure about the zeroing Hellburner? Because if it is not zeroed, we will have no hope of catching up with BbR

 

[Adul]

Originally posted by: UlricT

Originally posted by: Hellburner
All the beta numbers will be zeroed out once S@H2 actually starts.

are you sure about the zeroing Hellburner? Because if it is not zeroed, we will have no hope of catching up with BbR

Never say never

 

[Spacehead]

Originally posted by: UlricT

Originally posted by: Hellburner
All the beta numbers will be zeroed out once S@H2 actually starts.

are you sure about the zeroing Hellburner? Because if it is not zeroed, we will have no hope of catching up with BbR

I don't even think they're crunching real data, last i knew. Maybe they are now???

Originally posted by: SNC

Seti 2 I think will be a flop.

Why do you think that?

 

[SNC]

Originally posted by: SNC

Seti 2 I think will be a flop.

Why do you think that?[/quote]


Because of the lack of 3rd party apps to begin with.
2ndly the auto update of the software is going to be a real problem for some of the corprate users.
I will not allow the software to be installed on any of my office PCs untill a number of questions are answered.
Just my little world has 60 PCs and 5 servers. That accounts for around 275 WUs aday.
If there are 10 people like me that do not move to Boink that is 2750 WUs aday lost.

with around 10,000 aday being done by the TeAm, That is over 25%. I would call that a big flop. Wouldnt you?

 

[Thoreau]

Originally posted by: SNC

Originally posted by: SNC

Seti 2 I think will be a flop.

Why do you think that?

Because of the lack of 3rd party apps to begin with.
2ndly the auto update of the software is going to be a real problem for some of the corprate users.
I will not allow the software to be installed on any of my office PCs untill a number of questions are answered.
Just my little world has 60 PCs and 5 servers. That accounts for around 275 WUs aday.
If there are 10 people like me that do not move to Boink that is 2750 WUs aday lost.

with around 10,000 aday being done by the TeAm, That is over 25%. I would call that a big flop. Wouldnt you?[/quote]

Actually, we did 55k yesterday, with no less than 20k in the last four days alone. =) The new SETI isn't based off of Work Units, if I understand it correctly, but on actual CPU power donated (or something along those lines.) Either way, we *WILL* slam BBR there too, one way or another.

Plus, i'm sure that there will be more information and applications for the new system as time goes on. You don't think the SETI community would settle for just one client with no add-ons do ya?

 

[Wiz]

I ran the beta for 24 hours, lost a days production for seti 1 on my main machine and the credit for 75% of my work still hasn't been granted more than ten days later.
I fully understand it's a beta, and those in power don't care that much about credit given - however they can't afford to alienate people like us and the lack of credit given for work performed does just exactly that with me.
I also experienced weird problems with Boinc running, and I didn't see any bug report avenue that was easily accessable. So I ran it for 24 hours and that was enough for me.


Oh, another thing was the fact I couldn't see how to only do seti other than dumping work that I didn't care to do.
There should be some setting you can control to set what kind of work you want to do and what you don't want to do and it should be easily accessable.

 

[SNC]

Actually, we did 55k yesterday, with no less than 20k in the last four days alone. =) The new SETI isn't based off of Work Units, if I understand it correctly, but on actual CPU power donated (or something along those lines.)

That is not normal production numbers. A ton of that is from new users with a lot of WUs joining up.
The production is still 10-15K My example still stands, be it 25% or 20% or even 15%, that is a lot of work (units) howerver it is credited. I hope you are right about 3rd party apps. I really want to do Seti2. Please dont get me wrong. Just not as it is.

 

[Crazee]

When Seti first started there weren't any 3rd party apps either. Until BOINC is out for awhile, people will not really know what they need a 3rd party app for or if they do what they won't know how it should be written yet. Give it time. Seti didn't become the smooth running DC project it is today overnight

 

[Rattledagger]

Looks like someone didn't see what beta they signed up for, it was BOINC BETA TEST, not seti-for-boinc.

Since it clearly states "Computational results are not being recorded." of course the only you can expect is that they maybe saves a top-200-beta-crunchers-list, just like they did for seti@home... Also expect the current stats to be zeroed before they starts testing a real verifier...


As for auto-update, you can choose to not download new exe before ok-ing this, uhm after they've ported this from unix that is.
Of course, you probably very soon will not have anything to crunch if not accepting a new exe...


For credit, if following Wiz statement, of course it takes a long time to get any credit then 6/10 doesn't bother crunching their wu. And before the credits makes much sence anyway they must fix the current time-bug & program a real verifier... For every wu you're crunching and someone discards, you'll have to wait till the time-limit expires before it's being re-issued to someone else...


Except for non-networked-boxes, proxy should not really be a big problem, and AFAIK the only not yet included or planned included in BOINC is something like setiwatch to monitor multiple clients/machines. Then again, I've crunched a couple genome-wu, and didn't have any problem following the progress without such a tool...

 

[Rattledagger]

Well, for the lack of 3rd-party-tools, an alpha-version for BoincView is out, this can be used to monitor progress of networked machines.

As already mentioned in the thread, wrongly detects 2nd copy on multi-cpu as "ready to run"... Already finnished wu is correctly displayed.

There's 3 different tabs:
1: Host-tab: host-name -- ip -- cpu -- OS -- platform -- the 3 benchmarks & CS per hour -- client-version
2: Results-tab: wu -- application & version -- host-name -- cpu-time -- %done -- 3 different choises for end-time -- estimated credit -- deadline -- Status
3: Options-tab: Is used to add/remove hosts, set update-interval & choose the completion-time-displaying-method.
For update-interval, there's no point setting this lower than the corresponding BOINC-preference-setting on the web-page. This is default 60 seconds.

The 2nd tab can choose if wants to look on individual host-stats or for all hosts.
All columns on host-tab & results-tab can be sorted increasing/decreasing fashion, but it looks like atleast the MFlops isn't correctly sorted...

The 3 different choises for end-time-displaying is:
Time to completion, final cpu-time, date & time wu will be finnished.

 

[Idoxash]

I was running the astro beta yesterday "boinc" and it gave me 16 hrs for a work unit! I know my computer not fast but 16 hrs? Is that some sort of a bug or is it really going to take that long to do work units?

--Idoxash

 

[Rattledagger]

The 1MB big Astropulse-wu takes 2x the time of a seti-wu under BOINC. At the same time a seti-wu under BOINC uses 60% or something longer than v3.03 under seti@home, so Astropulse is nearly 3x compared to seti@home...


There's also currently a bug there the cpu-time of a running wu increases every time BOINC connects to the scheduling-server... If you lands on nearly the same as the "expected" time you got then downloading, it's not this bug.

 

[Wiz]

Rattledagger, I do a lot of beta testing. Generally there is a desire by the people creating the software to get bug reports.

I found what I feel is a significant bug, one that will keep me from running the beta on my machines, and I see no good avenue to report it.

That's reason #1 that I will not be running it again for the forseeable future.

Reason #2 is the very long delay between doing work and getting "credit" for it.
Who cares about the "credit" while it's in beta, except that it also seems to indicate that the project has some issues and the project leaders don't seem to care about them.

For me, forget it.

I originally started seti in May of 1999, I was signed up for it in the fall of 1998.

I won't run BOINC until they can figure things out better than they have now.
Just my choice of what I do with my cpu cycles

 

[Rattledagger]

For reporting bugs, either a post in the forum or to the sourceforge-beta-list.

For credit-time, the beta-test is currently not a good indication here due to bugs, not many participants and no validator.
Later most wu should be credited after 2 days, since most will not change the default profile of 1 to 3 days cached work.
From the bug/feature-list #996 is Pending credit, or they'll list the credit for not yet validated/invalidated results... So validated + pending is your up-to-date-score and should be usable in races and such things.

The only you will not get credit for is invalid results, but AFAIK over 99% of all invalid seti-results is either too much overclocking or cheaters, so not getting credit for this botch-up work shouldn't be a big problem anyway... :evil:

 

[SoulAssassin]

Originally posted by: SNC

Originally posted by: SNC

Seti 2 I think will be a flop.

Why do you think that?

Because of the lack of 3rd party apps to begin with.[/quote]

I'm sure S@H2 will begin w just as many 3rd party apps as the original started with. These things take time, I predict a slow start for s@h2 coinciding w a big rush for S@H1 once the final date is announced, followed by a small transition to s@h2 and a large transition to other projects, followed by a transition by a good portion of those people back to s@h2 as these apps begin to emerge and the bugs get worked out. By no means will it be a flop though. Consider how many of the millions of users are small people who are running it on a single machine (as 99% of us started). These people don't use apps like SETIDriver and SETIQueue and the fact that this version will actually cache some w/u's is nothing short of a miracle to them. When it comes to # of users, people like us are the minority. There's even the possibility that BOINC could evolve into the base client for most DC apps.