Viewing some of the other x@home projects, while surfing the Internet, I saw a graphic that was both informative and appealing, that I would like the E@H staff to consider appending to the “Server Status” page.
The graphic was a horizontal bar graph, which comprised of four small, differently colored bars, depicting the following information:
Guys, I fully understand there’s a lot going on, for your group at the present, but I’m just presenting something I’d like to see and perhaps be of use, in illustrating to new E@H’ers that credit is rendered only after various stages of validation, and not directly after a work unit has been “user” processed and uploaded.
Keep up the good work,
TA
Copyright © 2024 Einstein@Home. All rights reserved.
Graphic for Work Units in Validation Stages
)
Please, can you provide a URL? I'm not sure what's meant by the different stages of validation.
Bruce
Director, Einstein@Home
RE: Please, can you provide
)
Bruce – A many thanks for your honored response, but after spending the last 90 minutes of going through my “Favorites” folder, I cannot locate the URL in question, to illustrate my request.
However, perhaps I can explain what I “Think” I saw several weeks ago in a more explanatory manner, but please bear with me, since even though I have read BOINC’s WIKI concerning validation, I could be misinterpreting the intended meaning.
What I understand is; each work unit is processed a minimum of three times before it is accepted and categorized as a canonical resultant unit. Furthermore, different E@H subscribers may simultaneously be processing the same work unit and submitting their efforts to E@H’s validator, for pending credit; hence receiving credit, once all three work units have been processed and compared to a standard set of final acceptance criteria.
OK, if the aforementioned assumption, that I have outlined, is somewhat correct; then I have to refer back to my original posting of:
• Work Units in their 1st Stage of Validation (red)
• Work Units in their 2nd Stage of Validation (green)
• Work Units in their 3rd Stage of Validation (blue)
My sincere apologies for not being able to locate the needed URL, but hopefully this may add some clarification to my first posting.
Keep Up the Good Work,
Tom
Theory of Gravitational Waves & LIGO
Laser Interferometer Space Antenna - LISA
JPL-Caltech
In BOINC, the validation
)
In BOINC, the validation process does not start until enough WUs are on hand to do the validation. So, there is really only one stage of validation.
BOINC WIKI
RE: In BOINC, the
)
John – Thanks for the informative correction to my conception of how the process of validation is achieved. Without knowing or understanding the time element, of how long the validator takes to process a work unit computed three to four times, before credit is distributed to individuals, is it still possible to take a snap-shot in time of the total work units within the validator?
Please understand, what I am suggesting for consideration, may be a useless item of information, considering the “Server Status” page in updated on a 30 minute frequency. Also, I have again, for the third time, tried to locate the URL that has Bruce requested to illustrate the example “I Think I Saw” of a validation status indicator.
Bruce and John, thanks for your time and responses to my suggestion, should I be able to provide the aforementioned URL example, I certainly will provide a post within this thread.
Regards,
Tom
Theory of Gravitational Waves & LIGO
Laser Interferometer Space Antenna - LISA
JPL-Caltech
RE: RE: In BOINC, the
)
Each validator will have one WU at a time to validate. When the system is working correctly (S@H is having disk based delays) each validation should be fairly fast - on the order of seconds or less for most projects.
BOINC WIKI
RE: Each validator will
)
John – Thanks for the validation timing explanation, per work unit; it appears from your description that my suggestion would be impossible to implements and possibly only slow overall validation, by overhead factors.
Regards,
Tom
Theory of Gravitational Waves & LIGO
Laser Interferometer Space Antenna - LISA
JPL-Caltech
I don't know what you've seen
)
I don't know what you've seen either, but I can imagine three interpretations of "three stages of validation":
1. The numbers of Results that are
- "received but not validated" (should be very few on most projects)
- "validated by itself, awaiting quorum for the workunit"
- "assigned credit"
2. I can imagine a project where the validation process indeed takes a noticable time which might make sense to split up in three stages. However there is no project I know of to which this applies.
3. Assuming a quorum of three one might list the numbers of workunits which have one, two and three successful results returned.
BM
BM
Bernd & Guys – Before
)
Bernd & Guys – Before selecting a distributed processing project, we attempted to research as many different projects as we could locate; some were what we considered meaningful, others were difficult for us to comprehend their intended objectives, and relate our understanding of these projects to others who we wanted to solicit in participating with us (not as a team function but rather as a personal achievement exercise) in “a” (single) project.
I felt that I had “bookmarked” all uncovered projects, this is obviously not so, but for some odd reason I do remember, I’m fairly certain, a web page with a graphic at the top of the page, and above the graphic were the words “Units in Validation”. Now, understanding what I believe is reasonable to understand is what John posted:
Bernd, I also subscribe to what your posting outlines:
So, what did I view? I’m not certain now; or the web page was in error by a simple misunderstanding of what was intended by the Project’s Administrator to the web page’s programmer.
Bruce, John and Bernd, again thanks for all the understanding in my “Hey!, How about this” suggestion to bring perhaps a simpler approach to users like myself, to understand the “validation” process and why “credit” is not awarded nanoseconds after a work unit is uploaded.
Regards,
Tom
Theory of Gravitational Waves & LIGO
Laser Interferometer Space Antenna - LISA
JPL-Caltech
RE: I don't know what
)
validated by itself does not make sense
#3 looks like an interesting stat. May I expand?
WUs sent with:
0 results received -
1 result received -
2 results received -
waiting validation -
BOINC WIKI
John, maybe we have a
)
John,
maybe we have a different understanding of "validation". AFAIK a Result coming in is first checked ("validated") for syntactic correctness etc. using "check_single" of the (project-dependent) validator. Then, when a quorum for the WU has come together, it is compared with the other results using "match_two". The "check_single" was what I meant when writing "validated by itself".
BM
BM