I have been noticing that the numbers shown on my account page in "pending credit" do not match the numbers from the list on "computers on this account".

To clarify, when I log in, then click account, then pending credit, there are the same six WU that have been there for some time(two weeks?). When I click on my team, then view computers, view tasks, pending, I get what seems to be an accurate result list.

Does anyone know how to fix this? Am I actually accomplishing anything? I would hate to put wear and tear on my machines if the results are not going where they need to go.

Thanks in advance--RB

---

There's nothing wrong, not that I've figured out why but they've made some changes.

There's a tread about it http://setiathome.berkeley.edu/forum_thread.php?id=63430

The 'pending credit' feature has been pulled in conjunction with the revised credit calculations. The list you can reach via the 'pending credit' link on your account page will only have entries until the 8th of March or so, which is when the server code was updated to no longer enter 'claimed credit' into the database (which then showed up on the pending list until it validated).

As you found, looking at the task list will give you the number of pending tasks and you can get a guestimate of how much each task might pay from the list of validated tasks.

You should be accomplishing as much as before (check your RAC).

In other words: Everything is fine, you are just witnessing the demise of a useful tool.

---

Carola
-------
I'm multilingual - I can misunderstand people in several languages!

In a way, it's possibly a good thing that people will be encouraged to look at their complete task list (it's just two lines below 'pending credit' on your account page, you don't have to go all the way round via team and computers).

There are much more powerful filters there, for viewing pending tasks by application type, and so on - the only things you lose are the (potentially) huge long list on a single page, which took a lot of server resources and bandwidth to show you: and the total at the bottom....

I would encourage people to use the task list, and while they're there click the filters for 'invalid' and 'error' as well - they can be quick indicators of developing problems on your computer(s), which might need your attention.

whats worse on the servers. showing me one page with a lot of data or 20 pages that I have to click "next 20" again and again. I would think the repeated requests would be worse

---

In a rich man's house there is no place to spit but his face.
Diogenes Of Sinope

whats worse on the servers. showing me one page with a lot of data or 20 pages that I have to click "next 20" again and again. I would think the repeated requests would be worse

Probably depends how persistent/obsessive you are ;-)

Old way: you make one click, the server sends everything (whether you want it or not). Remember how people found that above a certain list size, they got nothing at all? I bet the people doing that weren't, in general, going to check each individual task.

New way: your one click returns much fewer tasks, though more data on each one.

If you click again, and again, and again... (all the way to the end of the list), then the cumulative server load - but distributed over time - is probably much greater. But if you get bored, and stop after two or three, it'll be much less.

So, what's the average boredom threshhold of a SETI@Home user? There's probably a PhD in there somewhere, if anyone wants to go for it...

add that to my S@H wishlist. I'd like to have the option to change the number of WU's presented per page perhaps 20 30 and 50 per page.

---

In a rich man's house there is no place to spit but his face.
Diogenes Of Sinope

Thanks all. Yea, I miss the old way already. That's cool though, as long as nothing is wrong.

Thanks for the quick answers..RB

---

In a way, it's possibly a good thing that people will be encouraged to look at their complete task list (it's just two lines below 'pending credit' on your account page, you don't have to go all the way round via team and computers).

There are much more powerful filters there, for viewing pending tasks by application type, and so on - the only things you lose are the (potentially) huge long list on a single page, which took a lot of server resources and bandwidth to show you: and the total at the bottom....

I would encourage people to use the task list, and while they're there click the filters for 'invalid' and 'error' as well - they can be quick indicators of developing problems on your computer(s), which might need your attention.

That's the first thing I do, when Logged In, check my latest, validated and creditted, error or inconclusive result, hate the latter, cause it's usually
the hardest to track down.
And Compairing results, stock vs SS(S)E3x opitimized, anonymus CUDA FERMI vs stock FERMI.
Different CPU make, AMD & INTEL, capabillities, Whetsone FLOPS, L2 or L3 cache, memory type, DDR2 or DDR3, Speed, 2 or 3 Channels, amount in GByte, HDD's
O.S. Windows XP, VISTA and WIN 7, *NIX, FreeBSD, DARWIN, i.e.

Different GPU's, nVIDIA's 8000; 9000; 200; 400(FERMI) and 500(FERMI),
ATI HD4830; 4850; 4870; 5750*; 5770*; 5850; 5870 and 6000 series.
(*=Single Precision.)
May seem strange, but out of work for a too long a time, so I consider
'this' partissipating in Distributed Computing Projects, my current job, among other work, on PC mostly and ocasually building a Custom Build PC or repairing one.(9 out of 10, PC is a Trashcan and when vacuuming and blowing the thing, a Back Up of the DATA and mostly a reinstall of the O.S. also mostly a Windoze Flavor, after checking the HDD's and a few 'other'things', PSU, [color=blue]COOLING, heatsinks, fans
CD/DVD/Blue Ray reader and or writer.
And a few thousend 'other'possible problems as f.i. Trojans, which are used in a BOTNET for sending SPAM, or a DDOS Attack, serious Hacking,
f.i. in the Netherlands, for public transportation a Card with RFID, which is
already Hacked and the app. is available on the INTERNET and
even on a DVD, being part of an magazine on this subject.
The card is already in use and with help from an RFID reader/writer, you can add
" money " and travel for free.
(They already added the difference and is getting payed by the users?!)

---

Fred, do you have to use colors that are unreadable < consider color blindness >
this was unreadable here

"May seem strange, but out of work for a too long a time, so I consider
'this' partissipating in Distributed Computing Projects, my current job, among other work, on PC mostly and ocasually building a Custom Build PC or repairing one.(9 out of 10, PC is a Trashcan and when vacuuming and blowing the thing, a Back Up of the DATA and mostly a reinstall of the O.S. also mostly a Windoze Flavor, after checking the HDD's and a few 'other'things', PSU, [color=blue]"

I understand using colors to quote or try to make a point but not everyone can see all < or any > of the colors. If they are too close to matching the background they're unreadable.

I know this is off-topic, but...

As someone who is significantly colour blind, I simply never read Fred's posts.

Give me black and white.

---

Ditto

---

Tim

Well, I agree about the Color Full post, also typed it in a hurry, about the worst
possible combination you can think of, ;-)

Infact, couldn't hardly read it myself, but noticed this, while logging in a few hours later so too late to edit or get rid of the colors.

(I'm "color-blind" myself, that is green, red and brown, can sometimes 'look', the same, depending of luminance and contrast.)

Then, I'll keep it Black and White, so to speak.

By the way, thanks for the advice.

---

I would like a first page of just WU counts, so I won't have to go through the lists task-by-task (a bummer, as many of you know). After an Update, I just want to make sure that all the tasks I report get onto SOME list, since we no longer have Pending Credit changes to add to the account credit total changes for that warm-and-fuzzy feeling that nothing got lost...

---

I would like a first page of just WU counts, so I won't have to go through the lists task-by-task (a bummer, as many of you know). After an Update, I just want to make sure that all the tasks I report get onto SOME list, since we no longer have Pending Credit changes to add to the account credit total changes for that warm-and-fuzzy feeling that nothing got lost...

Since you just want a count. A much simpler way would just be to just go to the last page of your pending or valid tasks and look at the offset number. If your offset count is 1240 & you have 6 on the last page then you are looking at 1246.

To have a count of tasks waiting for you whenever you open a page that number would have to be check at a given interval. I imagine they don't do this as it is probably an expensive query to run. It's not just CPU time but ties up disk access time as well. I am sure the option of slow down all science db access to generate a number you may or may not look at was considered at one point.

Generating the numbers on demand when you select something means you would have to wait for the query to finish in a similar way the pending credit list work(s/ed). The BOINC Devs would probably be the one to implement this server side feature. Apparently the pending credit query has an issue that the mega crunchers found where it wouldn't show everything. Could have been a server side time constraint on how long the query was allowed to run.

---

SETI@home classic workunits: 93,865 CPU time: 863,447 hours
Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[

Since you just want a count. A much simpler way would just be to just go to the last page of your pending or valid tasks and look at the offset number. If your offset count is 1240 & you have 6 on the last page then you are looking at 1246.

To have a count of tasks waiting for you whenever you open a page that number would have to be check at a given interval. I imagine they don't do this as it is probably an expensive query to run. It's not just CPU time but ties up disk access time as well. I am sure the option of slow down all science db access to generate a number you may or may not look at was considered at one point.

Generating the numbers on demand when you select something means you would have to wait for the query to finish in a similar way the pending credit list work(s/ed). The BOINC Devs would probably be the one to implement this server side feature. Apparently the pending credit query has an issue that the mega crunchers found where it wouldn't show everything. Could have been a server side time constraint on how long the query was allowed to run.

Problem is that to get to the last page of pending or valid takes forever for me (and would be a huge load on the database, too). And i never heard of his "offset number" before - what exactly is it?

Why can't the servers maintain such counts as they move tasks from status A to status B? They are updating the DB anyway, so not that much more work. Then when queried, they'd just have to pull the counts (a handful of bytes of data) rather than zillions of tasks.

---

last page? of 1240 results. thats a bit of traveling isn't it.


@jravin The offset number goes up by 20 every page you look at starting with 40,60,80 etc. all you need to do is look at the last numbers in your address bar as you browse your WU's

---

In a rich man's house there is no place to spit but his face.
Diogenes Of Sinope

last page? of 1240 results. thats a bit of traveling isn't it.


@jravin The offset number goes up by 20 every page you look at starting with 40,60,80 etc. all you need to do is look at the last numbers in your address bar as you browse your WU's

But you still have to walk through x number of pages to get to the last offset. For example yesterday, I checked my in-progress tasks (5,840 / 20 = 292) pages.

But you still have to walk through x number of pages to get to the last offset. For example yesterday, I checked my in-progress tasks (5,840 / 20 = 292) pages.

No, just edit the offset number in your address bar and hit enter, within a short while you should be able to roughly guess what number to use.

---

Kevin

last page? of 1240 results. thats a bit of traveling isn't it.


@jravin The offset number goes up by 20 every page you look at starting with 40,60,80 etc. all you need to do is look at the last numbers in your address bar as you browse your WU's

But you still have to walk through x number of pages to get to the last offset. For example yesterday, I checked my in-progress tasks (5,840 / 20 = 292) pages.

Actually, no you don't.

I've just tried:

* All tasks
* Next 20 (to get the format right)
* type '00' (turns offset=20 into offset=2000) - still tasks
* overtype '3' (offset=3000) - no tasks to display, off end
* overtype '25' (offset=2500) - still off end
* overtype '2' (offset=2200) - ah, back in task land.

and so on. You home in very quickly, with very little typing and clicking. I've got 2342 tasks in total across all machines, all states.