New Crediting system: questions

Message boards : Number crunching : New Crediting system: questions

To post messages, you must log in.

Previous · 1 · 2 · 3 · 4 · 5 · 6 . . . 7 · Next

AuthorMessage
MattDavis
Avatar

Send message
Joined: 22 Sep 05
Posts: 206
Credit: 1,377,748
RAC: 0
Message 22916 - Posted: 18 Aug 2006, 6:30:48 UTC - in response to Message 22914.  

P.S. I found this very interesting link to an informative page.


Whose site is that? It looks and sounds official, yet its not the correct url.


https://boinc.bakerlab.org/rosetta/forum_thread.php?id=2121#22652
ID: 22916 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
zombie67 [MM]
Avatar

Send message
Joined: 11 Feb 06
Posts: 316
Credit: 6,589,590
RAC: 386
Message 22918 - Posted: 18 Aug 2006, 6:37:49 UTC - in response to Message 22916.  

P.S. I found this very interesting link to an informative page.


Whose site is that? It looks and sounds official, yet its not the correct url.


https://boinc.bakerlab.org/rosetta/forum_thread.php?id=2121#22652


This takes us to the other thread. Okay, then what? What are you trying to tell us? What are you trying to say?



Reno, NV
Team: SETI.USA
ID: 22918 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
MattDavis
Avatar

Send message
Joined: 22 Sep 05
Posts: 206
Credit: 1,377,748
RAC: 0
Message 22920 - Posted: 18 Aug 2006, 6:44:31 UTC - in response to Message 22918.  

P.S. I found this very interesting link to an informative page.


Whose site is that? It looks and sounds official, yet its not the correct url.


https://boinc.bakerlab.org/rosetta/forum_thread.php?id=2121#22652


This takes us to the other thread. Okay, then what? What are you trying to tell us? What are you trying to say?




I was showing you who posted the link originally.

You asked "Whose site is that?" and that's the response: the guy who made that post.

ID: 22920 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
zombie67 [MM]
Avatar

Send message
Joined: 11 Feb 06
Posts: 316
Credit: 6,589,590
RAC: 386
Message 22921 - Posted: 18 Aug 2006, 6:55:33 UTC - in response to Message 22920.  


I was showing you who posted the link originally.

You asked "Whose site is that?" and that's the response: the guy who made that post.


I'm confused. You're talking about Steve Cressman? It is the same guy who posted the link in this thread earlier too.


Reno, NV
Team: SETI.USA
ID: 22921 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
MattDavis
Avatar

Send message
Joined: 22 Sep 05
Posts: 206
Credit: 1,377,748
RAC: 0
Message 22924 - Posted: 18 Aug 2006, 7:00:00 UTC - in response to Message 22921.  


I was showing you who posted the link originally.

You asked "Whose site is that?" and that's the response: the guy who made that post.


I'm confused. You're talking about Steve Cressman? It is the same guy who posted the link in this thread earlier too.



Okay, one step at a time.

Click the link I gave.

It will go to a very specific post. The poster says he's the one that typed up the data page you're curious in.

I mean, it can't be that much clearer.
ID: 22924 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
zombie67 [MM]
Avatar

Send message
Joined: 11 Feb 06
Posts: 316
Credit: 6,589,590
RAC: 386
Message 22930 - Posted: 18 Aug 2006, 7:13:26 UTC - in response to Message 22924.  

Okay, one step at a time.

Click the link I gave.

It will go to a very specific post. The poster says he's the one that typed up the data page you're curious in.

I mean, it can't be that much clearer.


Sounds like it should be, bt it's not. It just takes me to the beginning of the thread...no specific post

https://boinc.bakerlab.org/rosetta/forum_thread.php?id=2121#22652

In fact, there is no such post as #22652 in that thread.

In any case, the real author just now posted about it in that thread.



Reno, NV
Team: SETI.USA
ID: 22930 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
MattDavis
Avatar

Send message
Joined: 22 Sep 05
Posts: 206
Credit: 1,377,748
RAC: 0
Message 22931 - Posted: 18 Aug 2006, 7:15:14 UTC

I see that specific post just fine. Who knows why you can't!
ID: 22931 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Saenger
Avatar

Send message
Joined: 19 Sep 05
Posts: 271
Credit: 824,883
RAC: 0
Message 22933 - Posted: 18 Aug 2006, 7:20:38 UTC - in response to Message 22931.  
Last modified: 18 Aug 2006, 7:22:01 UTC

I see that specific post just fine. Who knows why you can't!

I can't see it as well, as my forum settings don't allow it without further intervention. All I see is this:
Only the first post and the last 75 posts (of the 144 posts in this thread) are displayed.
Click here to also display the remaining posts.


You can change the value of 75 in your forum settings, but this is a way to reduce the server and connection load for long threads, and usually 75 (I think it's the default) is enough.

Edit:
speeling and a new bit of info: tralala just answered your question over there.
ID: 22933 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
zombie67 [MM]
Avatar

Send message
Joined: 11 Feb 06
Posts: 316
Credit: 6,589,590
RAC: 386
Message 22935 - Posted: 18 Aug 2006, 7:25:30 UTC - in response to Message 22933.  
Last modified: 18 Aug 2006, 7:29:16 UTC

You can change the value of 75 in your forum settings, but this is a way to reduce the server and connection load for long threads, and usually 75 (I think it's the default) is enough.


What a messed up preference setting! I told it to show 500, but it turns out that once I save it, it changed that to 50. No error message of course. So I changed the "If over" cell to 999. That fixed it.

Now I can see it! I wonder why only this one post was hidden?


Reno, NV
Team: SETI.USA
ID: 22935 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile dcdc

Send message
Joined: 3 Nov 05
Posts: 1829
Credit: 115,416,399
RAC: 53,241
Message 22936 - Posted: 18 Aug 2006, 7:28:49 UTC
Last modified: 18 Aug 2006, 7:29:16 UTC

ID: 22936 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Feet1st
Avatar

Send message
Joined: 30 Dec 05
Posts: 1755
Credit: 4,690,520
RAC: 0
Message 23033 - Posted: 18 Aug 2006, 15:42:14 UTC

tralala is a volunteer. He was attempting to create a description of the credit system based upon posts over on Ralph and information he had gathered. And wrote it up in a PHP page in hopes that it would make it easier for the project team to publish information about the new system. So, he created a page that has the look and feel of a Rosetta page, and the best text be could muster. Hopefully only minor revision is required by the project team. He was just trying to save them time, and facilitate the flow of a info. on the credit system. So, not offical word, but he's made every effort to verify his facts.
Add this signature to your EMail:
Running Microsoft's "System Idle Process" will never help cure cancer, AIDS nor Alzheimer's. But running Rosetta@home just might!
https://boinc.bakerlab.org/rosetta/
ID: 23033 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Biggles
Avatar

Send message
Joined: 22 Sep 05
Posts: 49
Credit: 102,114
RAC: 0
Message 23146 - Posted: 18 Aug 2006, 20:40:15 UTC - in response to Message 22879.  

both, but the granted credit is what gets reported to the stats sites.


I consider this a problem. The credit granting on Rosetta is way out of line with BOINC as a whole. A machine will earn far more credit per hour/day/week on Rosetta than on any other project if it uses an optimised client.

It's not technically cheating, but it's certainly not fair either.
ID: 23146 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Avi

Send message
Joined: 2 Aug 06
Posts: 58
Credit: 95,619
RAC: 0
Message 23147 - Posted: 18 Aug 2006, 20:41:58 UTC - in response to Message 23146.  

both, but the granted credit is what gets reported to the stats sites.


I consider this a problem. The credit granting on Rosetta is way out of line with BOINC as a whole. A machine will earn far more credit per hour/day/week on Rosetta than on any other project if it uses an optimised client.

It's not technically cheating, but it's certainly not fair either.


The new system is still being tweaked, after its done, I suppose we shall see.
ID: 23147 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Athlonheizer
Avatar

Send message
Joined: 19 Jun 06
Posts: 2
Credit: 213,458
RAC: 0
Message 23519 - Posted: 19 Aug 2006, 17:52:32 UTC

As long as the credit system did not develop and it not evidently is according to which criteria the credit is assigned (short Wus have partly substantially more credit than long Wus) I counts else where.

Athlon
ID: 23519 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
MikeMarsUK

Send message
Joined: 15 Jan 06
Posts: 121
Credit: 2,637,872
RAC: 0
Message 23522 - Posted: 19 Aug 2006, 17:57:18 UTC

It should average out over time - so you get roughly the same credits/hour regardless of whether you pick short or long WUs.

ID: 23522 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Christoph Jansen
Avatar

Send message
Joined: 6 Jun 06
Posts: 248
Credit: 267,153
RAC: 0
Message 23525 - Posted: 19 Aug 2006, 18:13:20 UTC

There is no need to go some other place as the credits you see beside your name are still assigned by the old system. The new credits are only shown so everybody knows they are there and for purposes of tuning the whole system before they become the ones that are reported to the statistics' sites.
So no reaseon to worry about that (at least not now).

Regards,

Christoph
ID: 23525 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Rebirther
Avatar

Send message
Joined: 17 Sep 05
Posts: 116
Credit: 41,315
RAC: 0
Message 23530 - Posted: 19 Aug 2006, 19:03:10 UTC

Minimum of given credits should be 7. CPDN, SAP, Riesel have more credits than rosetta. I have finished a WU in 3h and got only 14,50, not good at all.
ID: 23530 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
ritzl

Send message
Joined: 16 Apr 06
Posts: 3
Credit: 42,287
RAC: 0
Message 23533 - Posted: 19 Aug 2006, 19:11:43 UTC

Just saw this discussion 'cause I was wondering what the new column meant. Figured I'd throw in my two cents.

Here are two WU's. The first was selected because it seems to have generated credit at the same rate as I have been under the old rules. The second is mine, generated on a MPB G4 1.67 GHz running an optimized client. Both are indicative of the results being generated on both computers.

While I don't claim to know the inner workings of the credit system, I would like to point out that under the new rules, I am getting approximately 45% of the credit that the first computer is getting per CPU second. Since it is unclear to me whether or why my results are less credit worthy than another result, it seems that something is amiss here. It seems that since I have no control over the result, equivalent credit should be granted for equivalent valid processing time.

32951533 28551877 15 Aug 2006 22:51:13 UTC 19 Aug 2006 1:44:06 UTC Over Success Done 12,944.58 131.64 38.90

# random seed: 1919138
# cpu_run_time_pref: 14400
# DONE :: 1 starting structures built 6 (nstruct) times
# This process generated 6 decoys from 6 attempts
# 0 starting pdbs were skipped

----

33133122 28722850 17 Aug 2006 7:51:30 UTC 18 Aug 2006 2:06:15 UTC Over Success Done 28,620.45 287.31 32.50

# cpu_run_time_pref: 28800
# random seed: 1709970
# DONE :: 1 starting structures built 33 (nstruct) times
# This process generated 33 decoys from 33 attempts


BOINC :: Watchdog shutting down...
BOINC :: BOINC support services shutting down...

I am NOT trying to suggest that all structures are of the same complexity, nor that a decoy is a decoy is a decoy. I am only pointing out that on similar capability computers, I am processing twice as long and getting 90% of the credit (new rules).

I seem to remember that this was a problem for we mac users, and some of the folks here at Team MacNN put in a lot of effort to address this imbalance. Are the new rules again putting macs at a disadvantage? Or is there just some black box magic to the scoring?

The SETI folks seem to have done pretty well at equating flops to flops, cross platform, such that optimization comes from using each platform's inherrent processing strengths. Could you all try to do that here?


ID: 23533 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Christoph Jansen
Avatar

Send message
Joined: 6 Jun 06
Posts: 248
Credit: 267,153
RAC: 0
Message 23534 - Posted: 19 Aug 2006, 19:15:04 UTC - in response to Message 23530.  
Last modified: 19 Aug 2006, 19:22:28 UTC

Minimum of given credits should be 7. CPDN, SAP, Riesel have more credits than rosetta. I have finished a WU in 3h and got only 14,50, not good at all.


Yes, you are right, and most 3,2 GHz Intels get around 30 to 40 in 3h, sometimes even 50, so either the next few WUs will show that or there is some problem. That would be 10 to 12 per hour and per core, which is also what I get with an X2 3800 per core. I'd think the next two or three WUs will show that (at least it is what I constantly found when browsing through some machines).

[Edit] Well... 16.73 from your next WU is not really near to 30 or so... looks like I'll have to correct that with respect to these WUs. I'd be interested how it turns out on my next ones, I am still crunching the CASP refinement targets and have not received these new ones.[/Edit]

Regards,

Christop
ID: 23534 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Stefan
Avatar

Send message
Joined: 12 Feb 06
Posts: 5
Credit: 15,058
RAC: 0
Message 23536 - Posted: 19 Aug 2006, 19:23:56 UTC - in response to Message 23534.  

Weeeeeeeeeee...

So much for my RAC being over 400, its taking a nose dive...lol

I don't really care too much about the credits, they cause more trouble than help, but I think there should be a minimum amount based on computer speed, not flops or whatever it is now that determines it. I feel sorry for those machines who get 10 credits and take twice as long. They should get more...

Human Stupidity Is Infinite...

ID: 23536 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Previous · 1 · 2 · 3 · 4 · 5 · 6 . . . 7 · Next

Message boards : Number crunching : New Crediting system: questions



©2024 University of Washington
https://www.bakerlab.org