Message boards : Cafe LHC : How many seconds per point ?
Message board moderation

To post messages, you must log in.

AuthorMessage
Mark Reiss
Avatar

Send message
Joined: 1 Sep 04
Posts: 8
Credit: 11,170
RAC: 0
Message 12281 - Posted: 22 Jan 2006, 21:17:59 UTC

Hi:

Can anyone tell me how many seconds per point (on average) for LHC@Home?

Mark Reiss






ID: 12281 · Report as offensive     Reply Quote
borandi

Send message
Joined: 9 Nov 05
Posts: 45
Credit: 20,379
RAC: 0
Message 12329 - Posted: 23 Jan 2006, 21:17:48 UTC - in response to Message 12281.  

Hi:

Can anyone tell me how many seconds per point (on average) for LHC@Home?

Mark Reiss



Seeing as you have lots of credits, then you (should) know:

An average computer gets 100 credits/day.
This is general across all BOINC projects (as far as I know).

Therefore it depends on:

a) How many computers you have
b) How fast your computers are.

An average computer, I find, is around 800Mhz.

BOINC@Hull . xfire . mySpace

ID: 12329 · Report as offensive     Reply Quote
Profile Bird-Dog

Send message
Joined: 16 Dec 05
Posts: 18
Credit: 1,523,201
RAC: 0
Message 13128 - Posted: 24 Mar 2006, 23:48:02 UTC

m
ID: 13128 · Report as offensive     Reply Quote
Nuadormrac

Send message
Joined: 26 Sep 05
Posts: 85
Credit: 421,130
RAC: 0
Message 13284 - Posted: 9 Apr 2006, 5:03:35 UTC
Last modified: 9 Apr 2006, 5:09:04 UTC

You can do your own averages, but given one doesn't get claimed credit, but awarded, there really are too many variables to give a hard and fast number. It's just too affected by what other people claim, and where the average falls. This then gives the average result, that the closer to a high end comp someone has, the more often they'll receive greater then claimed credit, and more infrequently they'll receive less...

For instance, if someone is crunching on an FX60!, there are mighty few processors out there that will computer a given amount of work in less time then this monster. Course AMD might have come out with better, but it was the FX60! (and actually over-clocked) that Intel was useing as a baseline, in their own tests to show the sort of improvement Conroe would bring to the table when it's released...

If a person has a Pentium III or older, one will probably very seldom to never receive more credit then they claim... There's just no way to isolate one unknown, which would be accross the board and apply to everyone. Now on claimed credit, it would be possible to do the math. Take the length of time the WU ran in seconds, and divide by the awarded credit. The result should be secs/credit...

BTW, other things can affect it also. As many already know, Rosetta and RALPH (rosetta alpha) award claimed credit, and establish no quorum). Someone just noted over on the RALPH board that for the same amount of runtime, they noticed their Windows client was claiming more credit then their Linux client and was asking about that. Some were telling him it's more a BOINC thing then a RALPH thing, so don't try to figure that one out. It is possible (based on his observations though), that claimed credit might very some from one OS (say Windows) to another (like Linux)...
ID: 13284 · Report as offensive     Reply Quote

Message boards : Cafe LHC : How many seconds per point ?


©2024 CERN