Faster credit: 4 copies of each WU initially being sent out now

Message boards : Number crunching : Faster credit: 4 copies of each WU initially being sent out now
Message board moderation

To post messages, you must log in.

1 · 2 · 3 · 4 . . . 5 · Next

AuthorMessage
Divide Overflow
Volunteer tester
Avatar

Send message
Joined: 3 Apr 99
Posts: 365
Credit: 131,684
RAC: 0
United States
Message 57758 - Posted: 27 Dec 2004, 21:58:31 UTC
Last modified: 31 Dec 2004, 16:58:30 UTC

"December 27, 2004
We are experimenting with redundancy parameters. We will continue to send 3 copies of each workunit, but will validate results and grant credit when 2 (rather than 3) of them are returned. This should result in faster validation."

This brings up a number of new questions. If credit is granted on only 2 validated results, what are the new rules for awarding credit? With only two results, the high / low elimination - give the middle value won't work anymore. Will it be an average of the first two validated results? Inquiring minds....

ID: 57758 · Report as offensive
Ingleside
Volunteer developer

Send message
Joined: 4 Feb 03
Posts: 1546
Credit: 15,832,022
RAC: 13
Norway
Message 57761 - Posted: 27 Dec 2004, 22:08:32 UTC
Last modified: 27 Dec 2004, 22:10:50 UTC

The same rules that's always been used. ;)

If wu validated with only 2 results passing validation, the lowest claimed.
If wu validated with 3 or more results passing validation, remove highest & lowest and average the rest.

Any results validated after wu was validated is given the same credit as the others, no re-calculation is done.

ID: 57761 · Report as offensive
virex
Volunteer tester

Send message
Joined: 14 Jul 01
Posts: 17
Credit: 294,731
RAC: 0
United States
Message 57762 - Posted: 27 Dec 2004, 22:13:11 UTC - in response to Message 57761.  

> The same rules that's always been used. ;)
>
> If wu validated with only 2 results passing validation, the lowest claimed.
> If wu validated with 3 or more results passing validation, remove highest
> & lowest and average the rest.
>
> Any results validated after wu was validated is given the same credit as the
> others, no re-calculation is done.
>
>
>
soo this could end up resulting in lower granted credit....cause if you are validating with only two it will take the lowest....where as with 3 you would take the middle.....hmmmm
ID: 57762 · Report as offensive
Profile RossM
Volunteer tester

Send message
Joined: 5 Apr 02
Posts: 37
Credit: 36,921
RAC: 0
United Kingdom
Message 57763 - Posted: 27 Dec 2004, 22:18:34 UTC - in response to Message 57762.  


> soo this could end up resulting in lower granted credit....cause if you are
> validating with only two it will take the lowest....where as with 3 you would
> take the middle.....hmmmm
>

Yes but that is already the case for people with slower computers anyway as the result underneath shows. I put in a claim for 43 credits for most WU's but virtually everybody doing the other 2 WU's has a Pentium 4 HT and only claim 28 credits. So not going to make much difference i think.

http://setiweb.ssl.berkeley.edu/workunit.php?wuid=6645517



ID: 57763 · Report as offensive
virex
Volunteer tester

Send message
Joined: 14 Jul 01
Posts: 17
Credit: 294,731
RAC: 0
United States
Message 57766 - Posted: 27 Dec 2004, 22:23:00 UTC - in response to Message 57763.  

> Yes but that is already the case for people with slower computers anyway as
> the result underneath shows. I put in a claim for 43 credits for most WU's but
> virtually everybody doing the other 2 WU's has a Pentium 4 HT and only claim
> 28 credits. So not going to make much difference i think.

true.....i'm just worried this will become the norm since it only needs one low claim to get that
before you atleast needed two
ID: 57766 · Report as offensive
STE\/E
Volunteer tester

Send message
Joined: 29 Mar 03
Posts: 1137
Credit: 5,334,063
RAC: 0
United States
Message 57835 - Posted: 28 Dec 2004, 2:59:59 UTC
Last modified: 28 Dec 2004, 3:00:50 UTC

I just spent the last few Hours going through a lot of my Wu's that had been granted credit already and under the new format where they will grant the credit after the first 2 Results are turned in & I would have recieved less credit for 90% of the Wu's and at best only the credit I already recieved in the other 10% of the Wu's.

Not once did I run across one where I would have recieved more credit ...

ID: 57835 · Report as offensive
Profile Geek@Play
Volunteer tester
Avatar

Send message
Joined: 31 Jul 01
Posts: 2467
Credit: 86,146,931
RAC: 0
United States
Message 57838 - Posted: 28 Dec 2004, 3:10:16 UTC

I don't believe the validator is even running at the moment. I have many results with 3 returned and some with 2 returned. All are still pending credit of any kind.
Boinc....Boinc....Boinc....Boinc....
ID: 57838 · Report as offensive
EclipseHA

Send message
Joined: 28 Jul 99
Posts: 1018
Credit: 530,719
RAC: 0
United States
Message 57839 - Posted: 28 Dec 2004, 3:13:08 UTC - in response to Message 57835.  

If "the other cruncher" is a Linux box and you're on windows, you'll get around 1/2 the credit you claimed. There's also major weirdness on mono vs dual cpus doing the same WU. Maybe this will get fixed in Boinc 6.x! (or boinc 8.x)


> I just spent the last few Hours going through a lot of my Wu's that had been
> granted credit already and under the new format where they will grant the
> credit after the first 2 Results are turned in & I would have recieved
> less credit for 90% of the Wu's and at best only the credit I already recieved
> in the other 10% of the Wu's.
>
> Not once did I run across one where I would have recieved more credit ...
>
>
>
ID: 57839 · Report as offensive
Hans Dorn
Volunteer developer
Volunteer tester
Avatar

Send message
Joined: 3 Apr 99
Posts: 2262
Credit: 26,448,570
RAC: 0
Germany
Message 57843 - Posted: 28 Dec 2004, 3:30:01 UTC - in response to Message 57839.  

> If "the other cruncher" is a Linux box and you're on windows, you'll get
> around 1/2 the credit you claimed. There's also major weirdness on mono vs
> dual cpus doing the same WU. Maybe this will get fixed in Boinc 6.x! (or
> boinc 8.x)
>
Yep, and the linux cruncher will always get 1/2 credit. No 'doze boxes to help out anymore...

Regards Hans
ID: 57843 · Report as offensive
Profile kinnison
Avatar

Send message
Joined: 23 Oct 02
Posts: 107
Credit: 7,406,815
RAC: 7
United Kingdom
Message 57848 - Posted: 28 Dec 2004, 3:51:31 UTC

I've been getting approx 35-40 credits for all my wu's, occasionally 20-25 if the other 2 are using pentiums 4's or dual cpu things.
Now I'm going to get 20-25 almost almost all the time! This will have a major effect on my avg. credit.
I run an athlon xp2800 o/c to 2250MHz, it runs the boinc cpu bechmark in about 2000 double precision mips.


Not happy.
<img border="0" src="http://boinc.mundayweb.com/one/stats.php?userID=268&amp;prj=1&amp;trans=off" /><img border="0" src="http://boinc.mundayweb.com/one/stats.php?userID=268&amp;prj=4&amp;trans=off" />
ID: 57848 · Report as offensive
Profile MattDavis
Volunteer tester
Avatar

Send message
Joined: 11 Nov 99
Posts: 919
Credit: 934,161
RAC: 0
United States
Message 57877 - Posted: 28 Dec 2004, 5:08:36 UTC

You know, I love the people on this board, but sometimes you can be really dense :)

You complain that 3 validations takes to long, so the Seti people set it to 2. Now, you're mad that 2 shortchanges you on credits. Can't anything make you happy? :)
-----
ID: 57877 · Report as offensive
EclipseHA

Send message
Joined: 28 Jul 99
Posts: 1018
Credit: 530,719
RAC: 0
United States
Message 57887 - Posted: 28 Dec 2004, 5:34:02 UTC - in response to Message 57877.  

Think 1 WU = X Credits..(a constant) Everyone would be happy!

Don't tell me that "boinc wasn't designed to do that and it makes it all bougus.."
Because I will will use a real bonic project that does just that - CP

In CP, it doesn't matter what your benchmark is, or how long it takes a WU to process.. All "trickles" get the same number of credits, and as a result, all WU's get the same amount of credits...

In CP, you did X work, and got Y credit, if that work took 1 hour or 1 day! (be it on windows or linux!)

The new Seti scheme will really suck in the eyes of Windows crunchers, if the "other cruncher" is a Linux box!

The "difference" between claimed credits (on the same HW) between Linux and Windows has been known since before Seti/Boinc went live, but has ben a low priority for the developers, as it doesn't impact the science... Well, it's been over 6 months, and maybe they'll stop adding features and start fixing stuff like this! (If I heald my breath until they did it, I'd not only die, but my bones would turn to dust!) In the meantime, I'll crunch seti on Linux boxes, as that's the only way to always get the credits you claim!

> You know, I love the people on this board, but sometimes you can be really
> dense :)
>
> You complain that 3 validations takes to long, so the Seti people set it to 2.
> Now, you're mad that 2 shortchanges you on credits. Can't anything make you
> happy? :)
>
ID: 57887 · Report as offensive
Profile Keith Kennedy

Send message
Joined: 28 May 99
Posts: 149
Credit: 244,165
RAC: 0
United States
Message 57888 - Posted: 28 Dec 2004, 5:37:21 UTC

What's the big deal about how much credit you get? Are you going to buy a new computer with your credits?
As long as each person's credit is computed in the same way, it's no big deal!
ID: 57888 · Report as offensive
Hans Dorn
Volunteer developer
Volunteer tester
Avatar

Send message
Joined: 3 Apr 99
Posts: 2262
Credit: 26,448,570
RAC: 0
Germany
Message 57900 - Posted: 28 Dec 2004, 6:53:49 UTC - in response to Message 57888.  
Last modified: 28 Dec 2004, 6:55:14 UTC

> What's the big deal about how much credit you get? Are you going to buy a new
> computer with your credits?
> As long as each person's credit is computed in the same way, it's no big deal!
>

It's a big deal, because the rest of seti is working fine (more or less) :o)

I'd say, add the number of completed WUs to the seti stats, and the discussion will end.

Regards Hans

P.S:
'nuff said.
ID: 57900 · Report as offensive
Professor Desty Nova
Avatar

Send message
Joined: 17 May 99
Posts: 59
Credit: 579,918
RAC: 0
Portugal
Message 57924 - Posted: 28 Dec 2004, 8:56:52 UTC - in response to Message 57887.  

> The "difference" between claimed credits (on the same HW) between Linux and
> Windows has been known since before Seti/Boinc went live, but has ben a low
> priority for the developers, as it doesn't impact the science... Well, it's
> been over 6 months, and maybe they'll stop adding features and start fixing
> stuff like this! (If I heald my breath until they did it, I'd not only die,
> but my bones would turn to dust!) In the meantime, I'll crunch seti on Linux
> boxes, as that's the only way to always get the credits you claim!

This has been fixed in the coming 4.5x-4.6x? BOINC. The problem is the windows client that is inflating the benchmarks (something about compiling optimizations).
SETI@home classic workunits: 1,985 CPU time: 24,567 hours



Professor Desty Nova
Researching Karma the Hard Way
ID: 57924 · Report as offensive
wrzwaldo
Avatar

Send message
Joined: 16 Jul 00
Posts: 113
Credit: 1,073,284
RAC: 0
United States
Message 57942 - Posted: 28 Dec 2004, 9:50:03 UTC - in response to Message 57924.  

> This has been fixed in the coming 4.5x-4.6x? BOINC. The problem is the windows
> client that is inflating the benchmarks (something about compiling
> optimizations).
>


From what I have seen on my 2GHz XP box 4.57 has dropped the claimed credit about an average of 10 as compared to the 4.13 client.



<img src="http://boinc.mundayweb.com/seti2/stats.php?userID=2259&amp;team=off">
ID: 57942 · Report as offensive
Profile Roks

Send message
Joined: 20 Dec 02
Posts: 55
Credit: 137,776
RAC: 0
Slovenia
Message 57943 - Posted: 28 Dec 2004, 10:02:59 UTC

Well, my laptop always claims around 70? results, but mostly gets only about 20-30. What's the idea??? It takes him 6 hours to process 1 wu, and than it gets 1/3 work granted. I mean, give us the same work units, so we can really se the difference. Why would I get only 1/3 of claimed work granted? I think seti should grant you the credit that you did, not more or less. By seti classic it was better (by credists). You did your wu, you got credit. Now you do a lot of work for practicly no credit... SETI stuff, plz fix that somehow, but not the way you "fixed" it until now, because you only made it worst.
<img src="http://www.boincstats.com/stats/banner.php?cpid=d2319b8f0ad14565556d0ba45b64e779">
ID: 57943 · Report as offensive
wrzwaldo
Avatar

Send message
Joined: 16 Jul 00
Posts: 113
Credit: 1,073,284
RAC: 0
United States
Message 57944 - Posted: 28 Dec 2004, 10:12:58 UTC - in response to Message 57943.  
Last modified: 28 Dec 2004, 10:48:41 UTC

> Well, my laptop always claims around 70? results, but mostly gets only about
> 20-30. What's the idea??? It takes him 6 hours to process 1 wu, and than it
> gets 1/3 work granted. I mean, give us the same work units, so we can really
> se the difference. Why would I get only 1/3 of claimed work granted? I think
> seti should grant you the credit that you did, not more or less. By seti
> classic it was better (by credists). You did your wu, you got credit. Now you
> do a lot of work for practicly no credit... SETI stuff, plz fix that somehow,
> but not the way you "fixed" it until now, because you only made it worst.
>

I would be interested to see what credit it claimed running the 4.57 client. I just looked at your laptop and it is really asking for an inflated credit. Why should your laptop get 60 more credits than a 3GHz intel when it (your laptop) takes almost twice as long to process the same WU?

http://setiweb.ssl.berkeley.edu/workunit.php?wuid=6583372
http://setiweb.ssl.berkeley.edu/workunit.php?wuid=6583369

In this case I think 1 credit for 1 WU would be a good thing. Then the faster processor would get the credit it was due.

And are you serious with this one???

http://setiweb.ssl.berkeley.edu/workunit.php?wuid=5889457

Your laptop took well over twice as long and asked for 111.17 credits! The other two asked for 26 and 29 credits. Yep the system needs fixed, but you're not going to be a happy camper if/when it does get fixed!
ID: 57944 · Report as offensive
Ned Slider

Send message
Joined: 12 Oct 01
Posts: 668
Credit: 4,375,315
RAC: 0
United Kingdom
Message 57949 - Posted: 28 Dec 2004, 11:14:10 UTC - in response to Message 57839.  

> If "the other cruncher" is a Linux box and you're on windows, you'll get
> around 1/2 the credit you claimed. There's also major weirdness on mono vs
> dual cpus doing the same WU. Maybe this will get fixed in Boinc 6.x! (or
> boinc 8.x)
>

If the Linux users checked out my website and downloaded optimized boinc clients for their machines, they'd actually claim slightly more credit than Windows users rather than the current half ;)

I spent a LOT of time optimizing these clients to level the playing field for the benefit of both Linux and Windows users, but people have got to use them. I've logged over a 1000 downloads to date but there must be many more linux users out there still not using them.

Ned


*** My Guide to Compiling Optimised BOINC and SETI Clients ***
*** Download Optimised BOINC and SETI Clients for Linux Here ***
ID: 57949 · Report as offensive
STE\/E
Volunteer tester

Send message
Joined: 29 Mar 03
Posts: 1137
Credit: 5,334,063
RAC: 0
United States
Message 57953 - Posted: 28 Dec 2004, 12:00:32 UTC
Last modified: 28 Dec 2004, 13:43:23 UTC

You complain that 3 validations takes to long, so the Seti people set it to 2. Now, you're mad that 2 shortchanges you on credits. Can't anything make you happy? :)
==========

I'm not complaining about anything Matt just stating what I found with my WU's, you didn't see me complaining about wanting my Credits faster anywhere on the board even though most of the time I have a ton of Pending WU's, I just take the Credits as they come & live with it.

I was perfectly happy with the 3 Result Validation system but it seems to me the Dev's caved in and reduced it to 2 now to pacify the people that are whining they want their credits faster.

Are we going to see the first Result returned getting Credit 6 months from now & all the other returned results for that WU getting the same Credit in order to speed it up even further ... ???

Maybe they are just trying to get in line with the rest of the Projects on the amount of Credit they give out Per WU, I've always felt the Seti Project gave out way more Credit Per WU than the rest of the Projects did by comparison.

I topped out at 4,101.89 RAC Credits just the other day here and I know with the same Computers over at the LHC Site I could only get up to about 2300-2400 RAC Credits running full time, so theres is a big difference to me between the Projects on the amount of Credit that is being issued for the amount of time spent crunching per WU...

No matter how the credit system is set up a certain percentage of the people are going to be unhappy with it and the rest happy with it. No matter how they do the Credit System somebody is going to get the short end of it & somebody is going to make out like a bandit with it. There are just simply to many different systems to make everybody happy...IMO

ID: 57953 · Report as offensive
1 · 2 · 3 · 4 . . . 5 · Next

Message boards : Number crunching : Faster credit: 4 copies of each WU initially being sent out now


 
©2024 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.