Hello Panther-X,
Some of your questions are covered in the Forum within HELP (upper-right). Click Help - Points (Points is near the bottom under the Overview column.)
https://secure.worldcommunitygrid.org/help/viewTopic.do?shortName=points
1) What exactly is the difference between points claimed and points credited? Your CPU claims points but the server adjusts this value and tries to grant everybody the same points for the same work. Always lots of controversy.

5) There is a huge difference in points. On my mobile app, I get 4,096 but when I log into the website, I get 29,123 which is a huge difference.
WCG started with a non-BOINC system. So 7 WCG points = 1 BOINC credit. This is a historical artifact that will disappear someday. Most of our programming effort goes into running projects.

11) I have read that the tasks for FightAIDS@Home are same for the mobile and a PC, I am curious as to how many points a PC gets per valid result.
I have an older PC that is slower than an i7. My last 3 validated FAHV units were granted 41.6, 45.1 and 39.4 WCG points. As stated in Help - Points, in theory a work unit is granted the same number of points no matter where it runs.

4) What is the purpose of "communication schedule"?
Most people do not set this. It is intended to help people who get cheaper data rates at night or on the weekend. I still make my personal calls mostly after 8 PM. Old habits die hard.


2) I have noticed that some of my completed tasks were either in "Pending" or "Project Backoff".
You are getting different Android messages than I get on my PC so my answer may be wrong. The recent changes on the server have caused some problems this month. The scheduler on the server seems to sometimes be overwhelmed. This has not happened to me for the last 2 days. When my computer tries to contact the scheduler and gets a back off message to defer communications until the scheduler is less busy, it will wait a short time, then try again. Each time the scheduler says it is still busy, it will wait a longer time before it tries again. After several work units are trapped in this loop, additional work units will not even try to communicate, but just wait until the earlier work units can communicate with the scheduler. I think that that is what "pending" is about. Maybe the staff have fixed the problem with the scheduler. Hope.

6) I joined WCG just over 10 days ago but on my Profile page, the Total Run Time is almost 3 times of that (0:034:13:00:44). How is this possible?
Each core can contribute 1 day of computing each day. This leads to a topic related to 3) and 7). You have good equipment, but the current generation of Android mobile devices have often turned out to be unable to deliver quad-core performance consistently. Some of our members discovered that their ROM code disabled ("too much") computing. Each device is different, but I personally recommend using 1 or 2 cores on a quad-core mobile. See if that eliminates your "Invalid" results.

9) Since I have additional tasks automatically downloaded on my cell waiting to be crunched, am I being penalized for that or not?
That is fine. I have my cache set to 0.1 days with no additional buffer. This downloads work units 2.4 hours before they will start running. This keeps me from downloading dozens of work units when the BOINC algorithm suddenly changes my "Duration Correction Factor" because a work unit terminates surprisingly early and BOINC decides I have a speedy supercomputer. I could change to 0.01 days and download a quarter of an hour before I need to, but that would be very trusting. What if the scheduler backs me off?


10) . . . . How do I successfully finish the current tasks and quit the app to avoid downloading any new tasks or discarding the current ones?
You want to set the WCG Project to "No New Tasks". I am uncertain how to do this on an Android device. 3 taps on the lower right screen? Exit BOINC after you have finished all your downloaded work, but that is not supposed to be necessary. Once all your work is computed and sent back you can format, update and spindle the app as you like.

"No Response"? Maybe related to the recent Scheduler problem. I am unfamiliar with the new Android messages.

Lawrence

On Q1, WCG is the only BOINC project on server v700+ still maintaining this [as a legacy hack]. It's going away... end of that part of the confusion.

If you want to know the raw credits a device is locally rated, peak performance [before the 7x points adjustment for another WCG legacy hack], run the benchmark, add up Whetstone+Dhrystone and divide [IIRC] by ~480 [think there's an Start Here FAQ discussing the calculation]. The result is the approximate credit a device 'claims' to be able to compute.

Thanks for the replies fellows.

lawrencehardin:
1) Thanks for the explanation. The server determines the points equally for all the same work. That's good I guess. Out of curiosity, is there a formula that the server uses to assign the final value?

5) Okay, it does make sense to concentrate on the advancement of the science. I read the points article that you linked and my initial impression was that it didn't apply to me since I was a new donor, thus didn't contribute with UD Windows. However, upon re-reading it, I got a better picture than before.

11) Okay, so while the points per task is the same, a faster system does more than a slower one so ends up with more completed tasks and thus, more points than a slower system. Good enough for me. Sorry to bother you but how long does it take? On my HTC, it takes roughly 12 hours per task and on my Samsung, takes roughly 24 hours. Both cells have the same exact usage.

4) Okay. That's weird since I don't remember doing any changes to it. Since I am connected to Wi-Fi 24/7, I guess it won't have any impact so will leave it as it is.

2) Yep, that makes sense to me and since then, I haven't seen it again. I do hope that the issue is permanently resolved.

6) Okay, so regardless of the speed of the core, each one contributes 1 day if crunched for 24 hours. Sounds weird to me since Core A with twice the speed of Core B would effectively be doing 2 days of work. Thanks for the tip, I have changed the number of CPUs from 4 to 3 on both devices and will see how many invalid results I get. BTW, is there any way to find out what these device names map to:
android_27e8762e -> 3 invalid results
android_5ffdb15f -> 2 invalid results
If so, it would make it very easy for me to troubleshoot which mobile is getting more invalid results. Also, is it normal to get non-zero points for the invalid results? I have gotten between 16.1 and 20.0 in the last update.

9) Thanks for that. I guess it is best to leave at default since I am not getting penalized for it. Am I slowing down science by caching these tasks or not? If yes, then I would like to prevent it so how do I change the setting?


Coleslaw
10) Thanks for the guidance! I found it along with the other options.

3) I didn't think about the chargers. As an experiment, I have switched the cables between my Samsung and HTC and will see what the results are. Moreover, i have shifted down from 4 to 3 CPUs in order to eliminate the invalid results.


SekeRob
1) Okay, so in the future, it will be removed and the points will return back to "normal".

Thanks for that. I did search and found that this link (https://boinc.berkeley.edu/flops.php) gives 404. Regarding the points to GigaFlops in the FAQs, which value of points is correct:
32289 / 700 = ~46.13 GigaFlops
4612 / 700 = ~6.59 GigaFlops (My guess is this one as it is just 2 cellphones)


12) When I changed the number of CPUs from 4 to 3, I expected that the task with the lowest completion percentage be stopped. My Samsung lived up to my expectations but not my HTC as shown below. Maybe this can be fixed in the next update:
http://s11.postimg.org/5v9oneow3/Paused_Task.png

To answer your question about project backoff, that is when the work units couldn't be uploaded. This doesn't mean it is necessarily something on your side. There has been mention of a time or two recently where people were having trouble reporting. Just give it some time and it works itself out. Several things can cause it to happen. Pending just means it is waiting to be sent, started, etc... where as project backoff means that it can't report in for the time being and is "backing off" or waiting for a while longer before making another attempt to report in. I hope that helps clear things up.

Thanks for the welcome and explanations provided, Coleslaw

I couldn't find any data posted by Kevin/knreed in that thread. I guess, I wouldn't be happy with the numbers when comparing my cell to a top-end i7

BTW, while I have already joined Crunching@EVGA, is it possible for me to set my Samsung to [H]ard|OCP while leaving my HTC on Crunching@EVGA? If so, how would that be done?