| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
| World Community Grid Forums
|
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 7
|
|
| Author |
|
|
Hypernova
Master Cruncher Audaces Fortuna Juvat ! Vaud - Switzerland Joined: Dec 16, 2008 Post Count: 1908 Status: Offline Project Badges:
|
If I go into the BOINC Stats and look for my user stats, I see that on the WCG project I have a ranking of 2'997. On the WCG ranking I am 3'277. That's a world of difference in my mad race to stardom.
----------------------------------------![]() I am not serious here, but I was wondering what makes the difference. Maybe the conversion factor between WCG credits and BOINC credits. If I calculate the conversion factor from what is credited in WCG and what is reported in BOINC I get the value of 6.999079 which is very near to 7. ![]() |
||
|
|
roundup
Veteran Cruncher Switzerland Joined: Jul 25, 2006 Post Count: 843 Status: Offline Project Badges:
|
In the WCG statistics there are also the credits of the United Devices era included. Many crunchers collected those credits before WCG moved to BOINC. BOINCstats constists of your BOINC credits only. That makes the difference.
|
||
|
|
JmBoullier
Former Community Advisor Normandy - France Joined: Jan 26, 2007 Post Count: 3716 Status: Offline Project Badges:
|
If/when WCG makes the conversion of its points system which was envisaged all WCG points will be divided by 7 and the result will be passed to 3rd-party sites as WCG's new BOINC credits. Then things will be more clear and everybody will be happy.
----------------------------------------Or not, depending on how much work they did before switching to BOINC. In fact, although I have had one thread of a P4 crunching under UD for several months I just notice that my ranking is only slightly better under BOINCstats (1543) than under WCG (1561). Seems that the UD effect has been completely offset by the quad that I built later. Edit: Added "only" to "slightly better" to clarify things for Ingleside... ![]() ---------------------------------------- [Edit 1 times, last edit by JmBoullier at Jan 17, 2010 4:42:23 PM] |
||
|
|
Ingleside
Veteran Cruncher Norway Joined: Nov 19, 2005 Post Count: 974 Status: Offline Project Badges:
|
If/when WCG makes the conversion of its points system which was envisaged all WCG points will be divided by 7 and the result will be passed to 3rd-party sites as WCG's new BOINC credits. Then things will be more clear and everybody will be happy. Or not, depending on how much work they did before switching to BOINC. You can never get a fair conversion from UD-points to Cobblestones, since there's many differences between the systems. Still, a conversion that on average for most users would be more or less "correct" would be to divide WCG-points gotten with BOINC-client with 7, and divide WCG-points gotten with UD-agent with 12. The different conversion-factors comes because UD-agent only reported run-time while BOINC-client reports cpu-time, and would expect majority of UD-agent users didn't patch their agent to use more than 60% cpu-usage. In fact, although I have had one thread of a P4 crunching under UD for several months I just notice that my ranking is slightly better under BOINCstats (1543) than under WCG (1561). Seems that the UD effect has been completely offset by the quad that I built later. Of course it's better under BOINC-stats, since some of the WCG-users has never used the BOINC-client, and therefore isn't included in the BOINC-stats. ![]() ![]() "I make so many mistakes. But then just think of all the mistakes I don't make, although I might." |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
haha, then I propose to multiply the BOINC to WCG points first by 7/12th, so proper WCG rankings are established. When the factor 7 was created, it was Linux benchmark based, and the Windows inflationary factor was never incorporated, though the Windows share outstrips the Linux share... [remember that 55% thing, Ingleside?]. And then the BOINC inflation struck, when the BOINC Linux compiled client was tweaked to match the Windows
---------------------------------------- PS: It's technical issues, some mentioned by knreed, that has most likely had this not happen, yet.
WCG
----------------------------------------Please help to make the Forums an enjoyable experience for All! [Edit 1 times, last edit by Sekerob at Jan 17, 2010 2:39:16 PM] |
||
|
|
JmBoullier
Former Community Advisor Normandy - France Joined: Jan 26, 2007 Post Count: 3716 Status: Offline Project Badges:
|
In fact, although I have had one thread of a P4 crunching under UD for several months I just notice that my ranking is slightly better under BOINCstats (1543) than under WCG (1561). Seems that the UD effect has been completely offset by the quad that I built later. Of course it's better under BOINC-stats, since some of the WCG-users has never used the BOINC-client, and therefore isn't included in the BOINC-stats. ![]() OK, I should have written "only slightly better" since I consider 1543 and 1561 as equivalent. My point is that 2 years ago it was 10,821/8,864, last year still 1,844/1,705, and now almost the same both sides. ---------------------------------------- |
||
|
|
knreed
Former World Community Grid Tech Joined: Nov 8, 2004 Post Count: 4504 Status: Offline Project Badges:
|
Unfortunately, most of the stat tables (user level, team level, etc) store points as an integer field. Only the host table (the most basic level) stores points as a decimal field. This wasn't a big deal with the UD points, but it is much more important with BOINC sized credit. As a result we need to transform the tables to all store decimals instead of integers.
This means we then need to change a bunch of code in the system to use double's instead of long's. We will also need to take an outage on the website in order to make the change (we will reload the stats tables with decimal data instead of integer). This will take quite awhile to run (for example the user stats table which stores user level daily stats by project has 122,827,383 rows). |
||
|
|
|