Seti Enhanced Credit Fair?

Message boards : Number crunching : Seti Enhanced Credit Fair?
Message board moderation

To post messages, you must log in.

Previous · 1 . . . 17 · 18 · 19 · 20 · 21 · 22 · 23 · Next

AuthorMessage
1mp0£173
Volunteer tester

Send message
Joined: 3 Apr 99
Posts: 8423
Credit: 356,897
RAC: 0
United States
Message 312811 - Posted: 21 May 2006, 23:42:38 UTC - in response to Message 312809.  

Ned - that chart is for the entire SETI@Home project.

http://www.teamocuk.com/cprojectcred.php?p=SAH

RAC is cute, RAC is interesting, but I can probably go back a year and find messages saying that RAC is pretty meaningless.

I haven't bothered to look at how RAC is calculated.

The graph I posted from BOINCstats is credit, and the credit graph is pretty flat, meaning that the number of granted credits per day is pretty steady.
ID: 312811 · Report as offensive
Profile Steve @ SETI.USA
Avatar

Send message
Joined: 5 Sep 04
Posts: 189
Credit: 1,016,797
RAC: 0
United States
Message 312813 - Posted: 21 May 2006, 23:49:16 UTC - in response to Message 312811.  

RAC is cute, RAC is interesting, but I can probably go back a year and find messages saying that RAC is pretty meaningless.

I haven't bothered to look at how RAC is calculated.

The graph I posted from BOINCstats is credit, and the credit graph is pretty flat, meaning that the number of granted credits per day is pretty steady.


Ned - I think you just pretty much admitted that you don't have the facts to support your position.

RAC is based on GRANTED CREDITS!

http://www.setiusa.net
ID: 312813 · Report as offensive
1mp0£173
Volunteer tester

Send message
Joined: 3 Apr 99
Posts: 8423
Credit: 356,897
RAC: 0
United States
Message 312822 - Posted: 21 May 2006, 23:58:40 UTC - in response to Message 312813.  

RAC is cute, RAC is interesting, but I can probably go back a year and find messages saying that RAC is pretty meaningless.

I haven't bothered to look at how RAC is calculated.

The graph I posted from BOINCstats is credit, and the credit graph is pretty flat, meaning that the number of granted credits per day is pretty steady.


Ned - I think you just pretty much admitted that you don't have the facts to support your position.

RAC is based on GRANTED CREDITS!

RAC is based on Granted Credits in the same way Kilometers per hour is based on Kilometers. You can't ignore the time component.

RAC is in theory the rate at which credit is granted. In practice, it's pretty rough.

You are saying "When I looked at my speedometer it said 70 kilometers per hour so my one-hour daily commute is 70 kilometers" and I'm saying "dude, just read the odometer...."
ID: 312822 · Report as offensive
Profile Pappa
Volunteer tester
Avatar

Send message
Joined: 9 Jan 00
Posts: 2562
Credit: 12,301,681
RAC: 0
United States
Message 312825 - Posted: 22 May 2006, 0:07:07 UTC - in response to Message 312768.  

Hans

I have to agree with what you have said, right now for me the jury is not quite decided... I have machine(s) that were moved over from Beta to run standard for a while to get an idea of a baseline... With the mix of 4.18's and 5.12's it is difficult to get a true running average... This is also complicated with people stopping workunits that will take weeks into the future to get a quorum...

So for another two-three weeks I will not hazard a guess as to whether Eric managed to duplicate the standard application (credit wise) or not

So Patience is the key, right now... A good cross section of CPU types and OS types is required to really see if the final number was good...

I will note that my pending credits is running a bit high right now... But that is okay also, it will catch up...

Hello all,

*riesenseufzer*

I give up ... It's useless ...

Regards,
Carsten


Don't.

The graph Crunch3r posted shows a drop by less than 10%. This means that the overall transition to enhanced worked reasonably well.

The bigger drop for us "optimized" folks is a different affair.

Regards Hans


Please consider a Donation to the Seti Project.

ID: 312825 · Report as offensive
Profile Steve @ SETI.USA
Avatar

Send message
Joined: 5 Sep 04
Posts: 189
Credit: 1,016,797
RAC: 0
United States
Message 312827 - Posted: 22 May 2006, 0:09:40 UTC - in response to Message 312822.  

RAC is based on Granted Credits in the same way Kilometers per hour is based on Kilometers. You can't ignore the time component.

RAC is in theory the rate at which credit is granted. In practice, it's pretty rough.

You are saying "When I looked at my speedometer it said 70 kilometers per hour so my one-hour daily commute is 70 kilometers" and I'm saying "dude, just read the odometer...."


Ned - The only thing you got right here is that the time component is indeed critical.

RAC is average credit over a period of time. Using your example, looking at your speedometer is merely a snapshot - not an average.

Let me use your example...
If I went to work today and traveled 70 miles at 70 MPH, I averaged 70 MPH (RAC) for one day. If I drove 40 MPH on the second day, it took me nearly 2 hours to get to work, and my average speed (RAC) for the 2 days is now 55 MPH. yes, I traveled the same distance, but it took nearly twice as long the second day.

A RAC of 600,000 means a team has averaged that amount in granted credit per day. That team is therefore producing more work and credit than a team averaging 500,000 credits/day.

http://www.setiusa.net
ID: 312827 · Report as offensive
Profile Pappa
Volunteer tester
Avatar

Send message
Joined: 9 Jan 00
Posts: 2562
Credit: 12,301,681
RAC: 0
United States
Message 312828 - Posted: 22 May 2006, 0:13:04 UTC - in response to Message 312813.  

Steve I recall the discussion... The Big reason RAC can fluctuate so much is that the time period that RAC is calculated on is very short... Unlike BOINC Stats which uses a longer period to come up with Average Credit... There are peaks and valley's on a daily basis...

Things like Merging Hosts affect RAC, statistical averaging "smooths through" to take care of the small differences on a day to day basis...


RAC is cute, RAC is interesting, but I can probably go back a year and find messages saying that RAC is pretty meaningless.

I haven't bothered to look at how RAC is calculated.

The graph I posted from BOINCstats is credit, and the credit graph is pretty flat, meaning that the number of granted credits per day is pretty steady.


Ned - I think you just pretty much admitted that you don't have the facts to support your position.

RAC is based on GRANTED CREDITS!


Please consider a Donation to the Seti Project.

ID: 312828 · Report as offensive
Profile Steve @ SETI.USA
Avatar

Send message
Joined: 5 Sep 04
Posts: 189
Credit: 1,016,797
RAC: 0
United States
Message 312834 - Posted: 22 May 2006, 0:22:18 UTC

Pappa - I understand how RAC can fluctuate - it is, after all, a daily average calculation. I am also aware of how adverse variables can affect it.

Going back to Ned's example - he doesn't seem to understand that on one day you drive 70 miles at 70 mph, which means you would have driven 1,680 miles for that entire day. On the second day, yes you travel the same 70 miles at 40 MPH, but for the entire day you would only drive 960 miles.

http://www.setiusa.net
ID: 312834 · Report as offensive
Odysseus
Volunteer tester
Avatar

Send message
Joined: 26 Jul 99
Posts: 1808
Credit: 6,701,347
RAC: 6
Canada
Message 312835 - Posted: 22 May 2006, 0:22:23 UTC - in response to Message 312822.  
Last modified: 22 May 2006, 0:23:16 UTC

RAC is in theory the rate at which credit is granted. In practice, it's pretty rough.

The fact that if you quit crunching altogether, your RAC will remain constant from the point your last WU was validated, illustrates one of the roughest edges.

The formula is a fairly simple application of exponential decay; it’s given in the BOINC Wiki (I can’t be bothered to look up the page ATM, but will gladly do so for anyone who has trouble finding it). The essence of it is that each time you receive credit, the amount granted is averaged with your previous RAC depreciated at a rate of 50% per week.

Note that certain statistics sites use their own formulas; BOINCstats for example, on its summary pages—although it shows the official RAC as well there, and also in the tables.
ID: 312835 · Report as offensive
Profile zoom3+1=4
Volunteer tester
Avatar

Send message
Joined: 30 Nov 03
Posts: 66343
Credit: 55,293,173
RAC: 49
United States
Message 312839 - Posted: 22 May 2006, 0:30:38 UTC - in response to Message 312810.  
Last modified: 22 May 2006, 0:30:58 UTC

My position is that the WU's size doesn't really matter as long as It's not too excessive(Largest is about 4hrs:50mins or so on Joker3, On Joker2 I have 3 files that are HUGE and are at 7:00, 7:18 and 7:23 respectively[hr/min]), All I know is under the previous system I never got the full claimed credit(which is fine), the Granted Credit varied, At times It would be around 8 credits and other times It would be near 24 or so, So It wasn't consistent. Under the New credit scheme You get what the majority files for or so I think(Quorum). If a WU takes 4 hours You get whatever the WU is supposed to be worth, If It has the same value as a shorter WU You are effectively getting less per hour. So as someone who has the optimized Crunch3r app(5.12 SSE3 from Crunch3r), the 5.3.12.Tx36 client from Trux on top of the 5.3.16 Boinc install, I don't see myself as a cheater, We shall see what the drop is for us all, Hopefully something will be done. Right now I'm restricting My internet access to seti, Telling It I have less bandwidth so that maybe I'll get smaller files(If I have to I'll set It to near dial-up speeds), Like smaller than 2.5 hours per WU or so. In any case I'll wait, But not forever.

You can set your internet connection to whatever you want. It won't affect the "size" of the wu you get. All work units are roughly the ssame "size" as in filesize. The difference comes in when you look at the "angle range" of the work unit. Some angle ranges simply take longer to crunch than others. You can't tell the angle range by the file size as all wu's are the same size. When you are talking about setting your connection speed, it seems to me like you are asking to get physically smaller files but it doesn't work that way.


Well I didn't know how It worked. So I was mistaken on that.
Savoir-Faire is everywhere!
The T1 Trust, T1 Class 4-4-4-4 #5550, America's First HST

ID: 312839 · Report as offensive
Profile Lee Carre
Volunteer tester

Send message
Joined: 21 Apr 00
Posts: 1459
Credit: 58,485
RAC: 0
Channel Islands
Message 312901 - Posted: 22 May 2006, 1:31:47 UTC - in response to Message 307552.  

Just out of interest why do you need the return immediately option?

I like to keep track of my machines as they also function as low access rate data servers for a process that runs on one of my older machines, and I know (or did know) how long it takes each one of them to complete a WU. But I do not always have direct access to them or the local network.

I have the same need, but the servers are mission-critical, so I do not run anything on them that isn't strictly needed.

If they break, I don't eat.

I use Servers Alive for monitoring, it doesn't depend on the speed of a work unit.

I've also found UltraVNC to be quite handy.

Nagios is rather good
for those that are familiar with the term it's a sourceforge project
Want to search the BOINC Wiki, BOINCstats, or various BOINC forums from within firefox? Try the BOINC related Firefox Search Engines
ID: 312901 · Report as offensive
1mp0£173
Volunteer tester

Send message
Joined: 3 Apr 99
Posts: 8423
Credit: 356,897
RAC: 0
United States
Message 312905 - Posted: 22 May 2006, 1:38:21 UTC - in response to Message 312827.  

RAC is based on Granted Credits in the same way Kilometers per hour is based on Kilometers. You can't ignore the time component.

RAC is in theory the rate at which credit is granted. In practice, it's pretty rough.

You are saying "When I looked at my speedometer it said 70 kilometers per hour so my one-hour daily commute is 70 kilometers" and I'm saying "dude, just read the odometer...."


Ned - The only thing you got right here is that the time component is indeed critical.

RAC is average credit over a period of time. Using your example, looking at your speedometer is merely a snapshot - not an average.

Let me use your example...
If I went to work today and traveled 70 miles at 70 MPH, I averaged 70 MPH (RAC) for one day. If I drove 40 MPH on the second day, it took me nearly 2 hours to get to work, and my average speed (RAC) for the 2 days is now 55 MPH. yes, I traveled the same distance, but it took nearly twice as long the second day.

A RAC of 600,000 means a team has averaged that amount in granted credit per day. That team is therefore producing more work and credit than a team averaging 500,000 credits/day.


Yet, either way, your odometer would say that you had travelled 140 miles over that two day period.

Your "average speed" would be something closer to 2.92 miles per hour (140/48 hours), not 55 miles per hour -- you didn't actually move during 45 of those 48 hours.

Now, if we looked at the odometers on 180,000 cars every single day, and we saw that some had driven 300 miles, and some were not driven at all, then we'd know, on average, how the fleet of 180,000 cars were used.

... and if we found that those 180,000 cars drove 21 million miles on Tuesday, and they drove 21 million miles on Wednesday, and they drove 21 million miles on Thursday, well, that doesn't tell me anything about how far you drove, but on average, every car drives 116 miles per day.

This is May. If I go back to April 11th, and I see that, on average, the 180,000 cars drove 21 million miles on Tuesday the 11th, and Wednesday the 12th, and Thursday the 13th, it's pretty safe to say that driving habits in early April and late May are pretty much the same.

Except "cars" are active participants and "miles" are credits granted.

The graph from BOINCSTATS shows somewhere around 24 million credits granted on April 11th, and something about 27 million on May 20th. The graph is pretty flat for the past 60 days.
ID: 312905 · Report as offensive
Profile Steve @ SETI.USA
Avatar

Send message
Joined: 5 Sep 04
Posts: 189
Credit: 1,016,797
RAC: 0
United States
Message 312918 - Posted: 22 May 2006, 1:45:09 UTC

Ned - did you read this in a following post?

Going back to Ned's example - he doesn't seem to understand that on one day you drive 70 miles at 70 mph, which means you would have driven 1,680 miles for that entire day. On the second day, yes you travel the same 70 miles at 40 MPH, but for the entire day you would only drive 960 miles.


Look at how far you would drive in 24 hours at 70 MPH vs. 40 MPH.

This is really a simple concept.

http://www.setiusa.net
ID: 312918 · Report as offensive
W-K 666 Project Donor
Volunteer tester

Send message
Joined: 18 May 99
Posts: 19401
Credit: 40,757,560
RAC: 67
United Kingdom
Message 312944 - Posted: 22 May 2006, 2:02:56 UTC - in response to Message 312760.  

Genuine Intel (R) Pentium(R) M processor 1600MHz running Microsoft Windows XP Professional Edition, Service Pack 2, (05.01.2600.00)
CPU time = 79,523.20 sec., Claimed credit = 8.81 CS/hr., Granted credit = 2.56 CS/hr.

NOTE: This was right from WinterKnight’s second example (5.12 14,287s granted 56.46 = 14.23/hr wu id=78922770). So who’s the big bad credit whore and cheat now???

Franz

Well certainly not me, You can look at my results any time. My computers are NOT hidden. I gave links to back my claim where are your's. If those result is still there for viewing you will see they were both done with standard apps and standard client. With your's I have no idea, they could be real or they could be figures you invented.

For that one computer to one computer reference it just shows mine is faster than yours. My other computer,a P3, runs at about 4 credits/hour.

Andy
ID: 312944 · Report as offensive
1mp0£173
Volunteer tester

Send message
Joined: 3 Apr 99
Posts: 8423
Credit: 356,897
RAC: 0
United States
Message 312947 - Posted: 22 May 2006, 2:08:00 UTC - in response to Message 312918.  

Ned - did you read this in a following post?

Going back to Ned's example - he doesn't seem to understand that on one day you drive 70 miles at 70 mph, which means you would have driven 1,680 miles for that entire day. On the second day, yes you travel the same 70 miles at 40 MPH, but for the entire day you would only drive 960 miles.


Look at how far you would drive in 24 hours at 70 MPH vs. 40 MPH.

This is really a simple concept.

Yes, I read your entire post. Contrary to popular belief, I understood it.

Mine is also simple concept: across the entire fleet, the change in granted credit is minimal.

If RAC and Credits are proportional across the entire fleet then these two numbers should track.

If they don't, RAC doesn't really mean what we want it to mean.
ID: 312947 · Report as offensive
Profile SargeD@SETI.USA
Volunteer tester
Avatar

Send message
Joined: 24 Nov 02
Posts: 957
Credit: 3,848,754
RAC: 0
United States
Message 312971 - Posted: 22 May 2006, 2:53:59 UTC - in response to Message 312601.  


I was commenting on sideband's lack of observation, inability to understand what he reads and lack of understanding simple facts, and was making no comment on the levelness of the credits across the BOINC projects. If sideband was to use the right app he can for a short period until the end of S4, due to finish mid June, still keep his credits/time up.

Andy


That does not answer my question. If you get 30 credits for 45 minutes work on Einstein and way less than that here, then how has enhanced made the credit system of Seti equal to that of other projects? Weren't we told that was the purpose of changing the way credits were granted? Sounds to me like they should have gone up and not down.

ID: 312971 · Report as offensive
kevint
Volunteer tester

Send message
Joined: 17 May 99
Posts: 414
Credit: 11,680,240
RAC: 0
United States
Message 312991 - Posted: 22 May 2006, 3:57:16 UTC - in response to Message 312835.  

RAC is in theory the rate at which credit is granted. In practice, it's pretty rough.

The fact that if you quit crunching altogether, your RAC will remain constant from the point your last WU was validated, illustrates one of the roughest edges.



This is absolutly untrue - if you stop crunching your RAC or Recent Average Credit, will start to drop and will eventually reach a 0.


Go ahead, give it a shot, stop crunching for a few days, your constant RAC will constitantly be lower every day.
ID: 312991 · Report as offensive
1mp0£173
Volunteer tester

Send message
Joined: 3 Apr 99
Posts: 8423
Credit: 356,897
RAC: 0
United States
Message 313000 - Posted: 22 May 2006, 4:29:02 UTC - in response to Message 312991.  

RAC is in theory the rate at which credit is granted. In practice, it's pretty rough.

The fact that if you quit crunching altogether, your RAC will remain constant from the point your last WU was validated, illustrates one of the roughest edges.



This is absolutly untrue - if you stop crunching your RAC or Recent Average Credit, will start to drop and will eventually reach a 0.


Go ahead, give it a shot, stop crunching for a few days, your constant RAC will constitantly be lower every day.

Actually, it depends on where you look.

On the SETI site itself (or any project site) RAC does not change until the next report.

On the various stats sites, you'd have to ask the people who implemented the site.

I have a machine that was "burning in" and according to my stats on this site, it has a RAC of 38. It hasn't contacted SETI since 4/23.
ID: 313000 · Report as offensive
archae86

Send message
Joined: 31 Aug 99
Posts: 909
Credit: 1,582,816
RAC: 0
United States
Message 313002 - Posted: 22 May 2006, 4:32:52 UTC - in response to Message 312991.  

The fact that if you quit crunching altogether, your RAC will remain constant from the point your last WU was validated, illustrates one of the roughest edges.

This is absolutly untrue - if you stop crunching your RAC or Recent Average Credit, will start to drop and will eventually reach a 0.

Go ahead, give it a shot, stop crunching for a few days, your constant RAC will constitantly be lower every day.
That turns out not to be the case--for this question. The RAC will only keep dropping so long as your pending credit results keep validating. When your pending is gone, the drop stops.

And yes, I have recently tried it. I had a couple of old Pentium III's running at 3% SETI resource share when they converted to enhanced. Needless to say, their completion rate is glacial, and their previous pending long since went away. As part of my logging I've logged their SETI RAC's daily for some weeks. As of this morning, one had shown the exact same RAC for eleven consecutive daily readouts, and the other for ten. One of them went EDF today and completed a result. So at tomorrow's readout its RAC will move for the first time in over a week, and it will move down, as the derating of the past will outweigh the credit for the new result.

ID: 313002 · Report as offensive
Profile Jord
Volunteer tester
Avatar

Send message
Joined: 9 Jun 99
Posts: 15184
Credit: 4,362,181
RAC: 3
Netherlands
Message 313038 - Posted: 22 May 2006, 6:41:00 UTC

http://boinc-wiki.ath.cx/index.php?title=RAC for if you want to know the calculation and decay rates.
ID: 313038 · Report as offensive
Grant (SSSF)
Volunteer tester

Send message
Joined: 19 Aug 99
Posts: 13854
Credit: 208,696,464
RAC: 304
Australia
Message 313066 - Posted: 22 May 2006, 7:38:27 UTC - in response to Message 312772.  

In case you missed it, everyone's RAC is just beginning to rapidly drop.

And in case you missed it RAC is an even more meaningless number than Whetstones. But i'm sure that won't stop people from getting all worked up about it.
Grant
Darwin NT
ID: 313066 · Report as offensive
Previous · 1 . . . 17 · 18 · 19 · 20 · 21 · 22 · 23 · Next

Message boards : Number crunching : Seti Enhanced Credit Fair?


 
©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.