New psummary.html column: Kfactor

Moderators: Site Moderators, FAHC Science Team

Post Reply
smoking2000
Posts: 471
Joined: Mon Dec 03, 2007 6:20 am
Location: Amsterdam
Contact:
Contact smoking2000

New psummary.html column: Kfactor

Post by smoking2000 »

Since two hours ago, the psummary pages contain a new column "Kfactor" (although the table header is not all pages).

What is the meaning of the values in this column? And will this column remain in the psummary pages, or was this just a test or hickup?

I would like to know this, as would all other 3rd party devs (I presume) who have to update their psummary parsers :)
Top
ChasR
Posts: 402
Joined: Sun Dec 02, 2007 5:36 am
Location: Atlanta, GA

Re: New psummary.html column: Kfactor

Post by ChasR »

Could it be to calculate the -bigadv WUs? They have a k factor.
Image
Top
smoking2000
Posts: 471
Joined: Mon Dec 03, 2007 6:20 am
Location: Amsterdam
Contact:
Contact smoking2000

Re: New psummary.html column: Kfactor

Post by smoking2000 »

Heh, just after this post I read the following on the FAH blog:
We will be adding one new field to the psummary web pages (eg, http://fah-web.stanford.edu/psummary.html) to include a new part of FAH stats. We will give more info soon, but for now, please note the additional column on the far right.

For third party tools which parse these web pages, we have added the column at the far right in the hopes that this would not break current tools' parsing routines. Note that we are also looking in developing more systematic ways to interact with third party utilities, although that is still in the works. For now, we will give heads up to changes here in this blog.
Top
smoking2000
Posts: 471
Joined: Mon Dec 03, 2007 6:20 am
Location: Amsterdam
Contact:
Contact smoking2000

Re: New psummary.html column: Kfactor

Post by smoking2000 »

Since the psummary is being updated, would this be the right time for some improvement requests?

- Move the <META> tags to the <head>, and not before the <html> as it is now, to make the code W3C valid.
- Don't use two different content formats for the Kfactor column (one uses: <font size=-1>1</font>, the other uses only: 3.30)
Top
VijayPande
Pande Group Member
Posts: 2058
Joined: Fri Nov 30, 2007 6:25 am
Location: Stanford

Re: New psummary.html column: Kfactor

Post by VijayPande »

We'll add more info on Kfactor pretty soon -- we wanted to get out ahead of things with the infrastructure though. Good point regarding the cleanup -- I'll pass that request along.
Prof. Vijay Pande, PhD
Departments of Chemistry, Structural Biology, and Computer Science
Chair, Biophysics
Director, Folding@home Distributed Computing Project
Stanford University
Top
smoking2000
Posts: 471
Joined: Mon Dec 03, 2007 6:20 am
Location: Amsterdam
Contact:
Contact smoking2000

Re: New psummary.html column: Kfactor

Post by smoking2000 »

Thanks
Top
toTOW
Site Moderator
Posts: 6399
Joined: Sun Dec 02, 2007 10:38 am
Location: Bordeaux, France
Contact:
Contact toTOW

Re: New psummary.html column: Kfactor

Post by toTOW »

Since not all projects use this factor (yet ?), I guess it would be better to choose an escape character (like "-" or "NA") on projects that doesn't use the "BigAdv" formula instead of 1 (imho, no bonus scheme is different from bonus scheme with k=1) ... I guess it would help 3rd party developers generalize their code to all projects
Image

Folding@Home beta tester since 2002. Folding Forum moderator since July 2008.
Top
smoking2000
Posts: 471
Joined: Mon Dec 03, 2007 6:20 am
Location: Amsterdam
Contact:
Contact smoking2000

Re: New psummary.html column: Kfactor

Post by smoking2000 »

toTOW wrote:Since not all projects use this factor (yet ?), I guess it would be better to choose an escape character (like "-" or "NA") on projects that doesn't use the "BigAdv" formula instead of 1 (imho, no bonus scheme is different from bonus scheme with k=1) ... I guess it would help 3rd party developers generalize their code to all projects
Please don't. A factor of 1 is a valid factor too. Also the value in column is now either an integer (\d+) or floating point (\d+\.\d+), this is easy to parse with regular expressions. Not so if you mix in characters for string like "NA" or "-".

The current values can be used in code like: $some_number * $kfactor. Or to compare kfactors ($project_x_kfractor > $project_y_kfactor), this isn't possible when strings are also used in the kfactor column.
Top
toTOW
Site Moderator
Posts: 6399
Joined: Sun Dec 02, 2007 10:38 am
Location: Bordeaux, France
Contact:
Contact toTOW

Re: New psummary.html column: Kfactor

Post by toTOW »

I guess you didn't get my point : if some projects are not using the bonus scheme (and I guess only BigAdv projects actually use it), it's not the same as projects using the bonus scheme with k = 1 ... so you can't set the Kfactor column to 1 by default as it is currently done ...
Image

Folding@Home beta tester since 2002. Folding Forum moderator since July 2008.
Top
P5-133XL
Posts: 2948
Joined: Sun Dec 02, 2007 4:36 am
Hardware configuration: Machine #1:

Intel Q9450; 2x2GB=8GB Ram; Gigabyte GA-X48-DS4 Motherboard; PC Power and Cooling Q750 PS; 2x GTX 460; Windows Server 2008 X64 (SP1).

Machine #2:

Intel Q6600; 2x2GB=4GB Ram; Gigabyte GA-X48-DS4 Motherboard; PC Power and Cooling Q750 PS; 2x GTX 460 video card; Windows 7 X64.

Machine 3:

Dell Dimension 8400, 3.2GHz P4 4x512GB Ram, Video card GTX 460, Windows 7 X32

I am currently folding just on the 5x GTX 460's for aprox. 70K PPD
Location: Salem. OR USA

Re: New psummary.html column: Kfactor

Post by P5-133XL »

I agree that a parseable number is a much better alternative to fill the column than some invalid characters like "NA" which would require some form of exception treatment. While a k=factor of 1 still produces a bonus using the bigadv equations, the k-factor of 0 won't so that is probably a safe default for non bonus WU's.

Really, the programmers are very likely to already understand these types of concepts and I trust them to do what is necessary and best for the project. I would contend that this level of programming detail does not need to be discussed here, until or unless they mess up.
Image
Top
VijayPande
Pande Group Member
Posts: 2058
Joined: Fri Nov 30, 2007 6:25 am
Location: Stanford

Re: New psummary.html column: Kfactor

Post by VijayPande »

I'll pass the default value issue along as well. This is easy to change in the psummary code, so this is more of an issue of what makes the most sense.
Prof. Vijay Pande, PhD
Departments of Chemistry, Structural Biology, and Computer Science
Chair, Biophysics
Director, Folding@home Distributed Computing Project
Stanford University
Top
harlam357
Posts: 222
Joined: Fri Jun 27, 2008 11:03 pm
Location: Alabama - USA
Contact:
Contact harlam357

Re: New psummary.html column: Kfactor

Post by harlam357 »

Thank you for adding it to the right. That definitely saved my code, as it's pretty "dumb" in regards to the psummary parsing. Other than being rather inflexible, it's worked great so I haven't messed with it much. This change gives me a reason to overhaul and make it smarter. :)

Based on these formulas:

Code: Select all

Total points = base points * bonus factor

The bonus factor is computed based on the time from when our server issues the work unit to when it receives the work unit (WU_time), the time from work issuance to when the deadline would expire (deadline_time), the time from issuance to when the work unit times out and is marked for reissuance (timeout_time), and a constant factor k.
If WU_time > timeout time, bonus factor = 1.
If WU_time <= timeout time, bonus factor = sqrt(deadline_time * k / WU_time)
.

I feel that if a Project does not qualify for a bonus, then a k value of 0 should be assigned... since, as someone else mentioned, a k value of 1 would be valid for Bonus Projects.

The addition of this k value also give us (I think) everything we need to correctly calculate bonus PPD! :)

A few other questions (since I want to try and be as accurate as possible with the calculation):

- Is there anywhere we can find the timeout_time? Is this the final deadline? I'm under the impression that the current bigadv trial projects are 3 days (neither the preferred or final deadline), but I imagine this value will vary with new projects later on. It will be essential to know this value to determine when to set the bonus factor back to 1 and effectively remove all bonus PPD from the calculation.

- To follow on that, is the deadline_time preferred or final deadline... I assume the former.

If there's anywhere I can find an example calculation using available information, that would be most helpful.
Top
ChasR
Posts: 402
Joined: Sun Dec 02, 2007 5:36 am
Location: Atlanta, GA

Re: New psummary.html column: Kfactor

Post by ChasR »

timeout_time = preferred deadline
deadtime_time = final deadline

YHM
Image
Top
Post Reply

Return to “Discussions of General-FAH topics”