Message boards :
Number crunching :
Validation Pending
Message board moderation
Author | Message |
---|---|
Send message Joined: 4 Oct 13 Posts: 27 Credit: 2,301,681 RAC: 7,632 |
I have 70 completed work units in the "validation pending" state, some of them recent that I just completed, but most of them are from years ago. Just wondering about the validation process - when do new ones get validated, and why so I have so many that are still in this state. Thanks. |
Send message Joined: 7 Aug 04 Posts: 2184 Credit: 64,822,615 RAC: 5,275 |
Validation is a boinc thing that is not used by cpdn. However, the cpdn server has not been fixed one way or another so that this label doesn't confuse people on this project. In other words, it is nothing to worry about. |
Send message Joined: 4 Sep 04 Posts: 12 Credit: 8,018,591 RAC: 171 |
In other words, it is nothing to worry about. So, what is the relationship, if there is one, between these 'validation pending' units and our credit status?. My credit status is only updated, as far as I can see, when those units are validated (not to mention pendings from several years ago which I no longer consider). In another active thread regarding credit status there is a slight debate between science vs numbers, and despite I (and others) agree that science is the matter, the fact is that credit is our almost only feedback, not only on project life itself but in our contribution to it. As I'm stating there, I miss a more active participation of the moderators in this forum, I think we deserve it as contributors to the project. Best regards |
Send message Joined: 16 Jan 10 Posts: 1084 Credit: 7,743,223 RAC: 3,504 |
So, what is the relationship, if there is one, between these 'validation pending' units and our credit status?.None. My credit status is only updated, as far as I can see, when those units are validated (not to mention pendings from several years ago which I no longer consider).The credits are allocated when the credit script is run - there is no BOINC-based validation process, as validation on this project is performed by the scientists. From a BOINC perspective validation is a means of determining whether a new model should be released to complete a work unit and that is not done here: the condition for completing a work unit here is the successful completion of a single model. As I'm stating there, I miss a more active participation of the moderators in this forum, I think we deserve it as contributors to the project.The moderators have no ability to run the credit script, if that is what you mean: messages are passed onto the project if the approximately weekly credit script is significantly delayed. |
Send message Joined: 4 Oct 13 Posts: 27 Credit: 2,301,681 RAC: 7,632 |
Thanks, geophi. |
Send message Joined: 4 Sep 04 Posts: 12 Credit: 8,018,591 RAC: 171 |
The moderators have no ability to run the credit script, if that is what you mean: messages are passed onto the project if the approximately weekly credit script is significantly delayed. No, I didn't mean that at all, I wasn't even criticizing anything either. I was only asking for information about how the things work, for notices when there is a problem, for us, contributors, to understand it. For example, for the third weekend, this Sunday is ending again without credit updates. It may not be a significant delay yet. I'm not worried at all about my numbers. I'm just asking for understand these processes. And by the way, every time credit scripts are run, my validation pending is set to 0 (speaking only about recent tasks) so you may understand my surprise when you tells that the relationship is none, as I thought that that validation was related to confirming the completed and uploaded tasks, in order to reflect them in our personal credits. Anyway, thank you for informing. Best regards |
Send message Joined: 15 May 09 Posts: 4532 Credit: 18,746,213 RAC: 18,055 |
And by the way, every time credit scripts are run, my validation pending is set to 0 (speaking only about recent tasks) so you may understand my surprise when you tells that the relationship is none, as I thought that that validation was related to confirming the completed and uploaded tasks, in order to reflect them in our personal credits. The way I understand it is that the BOINC server code does that even if the project don't ask it to. I have no idea if there is a way to turn it off or not and reading the wiki for the server side of BOINC would probably just make my brain ache. |
Send message Joined: 31 Dec 07 Posts: 1152 Credit: 22,363,583 RAC: 5,022 |
Most projects use a validation process where they send the task to 2 or more people and wait until multiple results are returned to the project and then compare them. If the results match reasonably well then they are “validated.” This is fine for other projects who’s tasks compute in a few hours. Because of the length of CPDN tasks this is not practical. When CPDN was created 10 or more years ago running just one of the 160 year tasks to completion took 6 – 9 months. Waiting for 2 or 3 Crunchers to complete the same task could have taken (real) years. |
Send message Joined: 5 Aug 04 Posts: 1496 Credit: 95,522,203 RAC: 0 |
Building on JIM's post, which is spot-on: CPDN tasks are validated - with scientific algorithms. You can rest assured that scientists submitting work to CPDN are interested in reliable results. After all, numerous peer-reviewed papers have been published based upon our work (theirs and ours). "We have met the enemy and he is us." -- Pogo Greetings from coastal Washington state, the scenic US Pacific Northwest. |
Send message Joined: 25 Sep 09 Posts: 12 Credit: 370,712 RAC: 0 |
I have 72 WUs "waiting for validation". There is a recent one (one week), but other 71 are dated 2011-2013 and have credit. Why don't they change to "valid" status ? Are you still waiting for a confirmation ? |
Send message Joined: 15 May 09 Posts: 4532 Credit: 18,746,213 RAC: 18,055 |
You can ignore this. It is to do with the way the BOINC server code works. As noted below, validation is by statistical analysis of results and doesn't affect credit. However because the statistical analysis is independent of BOINC it doesn't send the server code a message to change the status. On other projects that validate when a second computer has the same result for the same work unit, BOINC then validates both and gives the credit at the same time. As you probably know, credit is based on trickle up messages for CPDN and has nothing to do with the unused validation system of BOINC. |
©2024 cpdn.org