Aaron, did you update the project from whose website you changed the preferences? Updating the website means that your client will have to call home to see the changes. Highlight the name of the project you made the changes on and then click "update" on the projects tab of the manager.

I.E if you updated the Rosetta website, then go to the Projects tab, hightlight the rosetta project (right hand box, click "rosetta" till it turns blue), then click "Update". Then look at Messages tab, you should see it.

Uhm... mmciastro, this is the RALPH@Home project... why would I be talking here about other projects settings?

I'm talking about changing from no setting to 2 hour setting on the Ralph@Home project preferences on this website we are on. Once I do this, and then click update in the BOINCMGR Projects tab with RALPH highlighted, it is supposed to get my new setting information. It appears to do so but does not tell me that my project preferences have been updated in the messages tab.

This is the message I get...

2/17/2006 5:13:43 PM||request_reschedule_cpus: project op
2/17/2006 5:13:44 PM|ralph@home|Sending scheduler request to https://ralph.bakerlab.org/ralph_cgi/cgi
2/17/2006 5:13:44 PM|ralph@home|Reason: Requested by user
2/17/2006 5:13:44 PM|ralph@home|Reporting 2 results
2/17/2006 5:13:49 PM|ralph@home|Scheduler request to https://ralph.bakerlab.org/ralph_cgi/cgi succeeded

In there, is no mention that I switched my RALPH@HOME settings (on the website preferences) to crunch work units from 2 to 4 hours. - Even though the change has clearly been made and is reflected by the amount of time that workunits are now taking. (it has jumped from 2 to 4 as expected).

---

you could always look in the client_state.xml file to see if the prefs have actually been updated, although this won't solve the original problem

At first, I didn't set this preference, & it appears that all of my workunits were processing for 1 hour or less. I have now changed my preference to 2 hour workunits, and clicked "Update", but it didn't give me any message in the "messages" tab of the BOINCMGR telling me that the preference had changed. Surely some kind of message letting me know that my preferences have changed should be there yeah?

Sorry Aaron, I misunderstood your statement above that I have placed in Bold. I read it to mean you edited the website general prefs and then clicked update at the bottom of the screen. I didn't see any mention of you updating from the "projects" tabs in the manager. Sorry.

just trying to help
tony

Machine 45 (5.2.13) is @ 4% after 2:00:06, looks right on target,
:)

Sorry Aaron, I misunderstood your statement above that I have placed in Bold. I read it to mean you edited the website general prefs and then clicked update at the bottom of the screen. I didn't see any mention of you updating from the "projects" tabs in the manager. Sorry.

just trying to help
tony

Oh, heh :) well I just figured you guys would assume all that of course.

I was just like.. this guy is confused? lol

---

Have I understood this discussion correctly? If I download WUs as before they will run a prediction and then end after a varaiable amount of time and return the results. If I set the CPU run time preference it will run sufficient predictions on the same WU to process up to that amount of time (Einstein runs different calculations on the same data, but does not have a time limit). If so what happens to the partial result? I.e. if I set it to 2 hours then I assume the work that is cut off at 2 hours will only be a partial result. Is this prediction ignored?

I believe the leave resdient in memory issue is fixed (from the other thread) so is this the only item currently worth testing on RALPH? I.e. should we all set such a CPU limit?

---

I have done a few results now. I haven't changed my preferences from day one. These are my General Prefs:

Processor usage
Do work while computer is running on batteries?
(matters only for portable computers) yes
Do work while computer is in use? yes
Do work only between the hours of (no restriction)
Leave applications in memory while preempted?
(suspended applications will consume swap space if 'yes') yes
Switch between applications every
(recommended: 60 minutes) 180 minutes
On multiprocessors, use at most 1 processors
Disk and memory usage
Use no more than 400 GB disk space
Leave at least
(Values smaller than 0.001 are ignored) .25 GB disk space free
Use no more than 85% of total disk space
Write to disk at most every 600 seconds
Use no more than 100% of total virtual memory
Network usage
Connect to network about every
(determines size of work cache; maximum 10 days) 3 days
Confirm before connecting to Internet?
(matters only if you have a modem, ISDN or VPN connection) no
Disconnect when done?
(matters only if you have a modem, ISDN or VPN connection) no
Maximum download rate: 200 KB/s
Maximum upload rate: 200 KB/s
Use network only between the hours of
Enforced by versions 4.46 and greater (no restriction)
Skip image file verification?
Check this ONLY if your Internet provider modifies image files (UMTS does this, for example).
Skipping verification reduces the security of BOINC. no

And these are my Ralph prefs:

Resource share
If you participate in multiple BOINC projects, this is the proportion of your resources used by RALPH@home 10
Percentage of CPU time used for graphics not selected
Number of frames per second for graphics not selected
Target CPU run time not selected
Miscellaneous
Should RALPH@home send you email newsletters? yes
Should RALPH@home show your computers on its web site? yes
Default computer location home
note: I don't have any seperately listed "home prefs" configured.

Host 103 (a celeron 500, 256M ram, and Win98se) has done these WUs:
3960 3872 17 Feb 2006 20:48:45 UTC 18 Feb 2006 3:19:28 UTC Over Success Done 3,219.00 2.15 2.15
3959 3871 17 Feb 2006 20:48:45 UTC 18 Feb 2006 3:19:28 UTC Over Success Done 2,464.00 1.65 1.65
3958 3870 17 Feb 2006 20:48:45 UTC 18 Feb 2006 3:19:28 UTC Over Success Done 2,128.00 1.42 1.42
3928 3840 17 Feb 2006 20:48:45 UTC 18 Feb 2006 12:44:08 UTC Over Success Done 2,092.00 1.40 1.40
3827 3739 17 Feb 2006 20:48:45 UTC 18 Feb 2006 7:59:05 UTC Over Success Done 4,136.00 2.76 2.76
3825 3737 17 Feb 2006 20:48:45 UTC 18 Feb 2006 3:23:37 UTC Over Success Done 8,003.00 5.35 5.35
2701 1610 17 Feb 2006 6:04:47 UTC 17 Feb 2006 7:28:34 UTC Over Success Done 900.00 0.60 0.60
2574 2516 17 Feb 2006 1:51:31 UTC 17 Feb 2006 7:11:50 UTC Over Success Done 643.00 0.43 0.43
2573 2515 17 Feb 2006 1:51:31 UTC 17 Feb 2006 7:03:30 UTC Over Success Done 1,410.00 0.94 0.94
2572 2514 17 Feb 2006 1:51:31 UTC 17 Feb 2006 6:59:21 UTC Over Success Done 4,807.00 3.21 3.21
2571 2513 17 Feb 2006 1:51:31 UTC 17 Feb 2006 3:41:48 UTC Over Success Done 803.00 0.54 0.54
2570 2512 17 Feb 2006 1:51:31 UTC 17 Feb 2006 3:41:48 UTC Over Success Done 1,355.00 0.91 0.91
2036 1984 16 Feb 2006 19:49:14 UTC 16 Feb 2006 22:39:19 UTC Over Success Done 1,951.00 1.30 1.30
1815 1780 16 Feb 2006 3:51:04 UTC 16 Feb 2006 20:38:33 UTC Over Success Done 895.00 0.60 0.60
1814 1779 16 Feb 2006 3:51:04 UTC 16 Feb 2006 20:26:06 UTC Over Success Done 894.00 0.60 0.60
1813 1778 16 Feb 2006 3:51:04 UTC 16 Feb 2006 20:10:15 UTC Over Success Done 4,173.00 2.79 2.79
1812 1777 16 Feb 2006 3:51:04 UTC 16 Feb 2006 19:40:54 UTC Over Success Done 581.00 0.39 0.39
1811 1776 16 Feb 2006 3:51:04 UTC 16 Feb 2006 14:05:37 UTC Over Success Done 719.00 0.39 0.39

Host 64, (and AMD 64 3700, Sandiego core, 939 socket, 1M L2, 1M ram, Winxpsp2) has done these:
1505 1470 16 Feb 2006 0:36:32 UTC 18 Feb 2006 2:54:52 UTC Over Success Done 3,577.66 14.13 14.13
1504 1469 16 Feb 2006 0:36:32 UTC 17 Feb 2006 11:33:01 UTC Over Success Done 3,645.92 14.40 14.40
1396 1361 16 Feb 2006 0:36:32 UTC 17 Feb 2006 11:33:01 UTC Over Success Done 3,552.34 14.03 14.03
1360 1325 16 Feb 2006 0:36:32 UTC 23 Feb 2006 0:36:32 UTC In Progress Unknown New --- --- ---
1359 1324 16 Feb 2006 0:36:32 UTC 18 Feb 2006 11:11:29 UTC Over Success Done 3,651.41 14.42 14.42

Host 66, (An AMD64 3700, 754 socket, 1M ram, Winxpsp2) has done:
1547 1512 16 Feb 2006 0:37:25 UTC 18 Feb 2006 3:45:47 UTC Over Success Done 3,505.02 13.11 13.11
1545 1510 16 Feb 2006 0:37:25 UTC 17 Feb 2006 19:36:24 UTC Over Success Done 3,517.22 13.15 13.15
1544 1509 16 Feb 2006 0:37:25 UTC 17 Feb 2006 19:36:24 UTC Over Success Done 3,653.16 13.66 13.66
1537 1502 16 Feb 2006 0:37:25 UTC 17 Feb 2006 6:25:07 UTC Over Success Done 3,637.28 13.18 13.18
1536 1501 16 Feb 2006 0:37:25 UTC 17 Feb 2006 19:36:24 UTC Over Success Done 3,538.94 13.24 13.24
1535 1500 16 Feb 2006 0:37:25 UTC 17 Feb 2006 14:49:33 UTC Over Success Done 3,576.41 12.96 12.96
1534 1499 16 Feb 2006 0:37:25 UTC 17 Feb 2006 14:49:33 UTC Over Success Done 3,809.81 13.80 13.80
1533 1498 16 Feb 2006 0:37:25 UTC 17 Feb 2006 11:34:10 UTC Over Success Done 3,527.30 12.78 12.78
1532 1497 16 Feb 2006 0:37:25 UTC 17 Feb 2006 11:22:37 UTC Over Success Done 3,624.30 13.13 13.13
1531 1496 16 Feb 2006 0:37:25 UTC 17 Feb 2006 11:22:37 UTC Over Success Done 3,592.52 13.02 13.02
1530 1495 16 Feb 2006 0:37:25 UTC 17 Feb 2006 11:22:37 UTC Over Success Done 3,523.59 12.77 12.77
1518 1483 16 Feb 2006 0:37:25 UTC 17 Feb 2006 6:25:07 UTC Over Success Done 3,511.05 12.72 12.72
1506 1471 16 Feb 2006 0:37:25 UTC 17 Feb 2006 11:22:37 UTC Over Success Done 3,743.23 13.56 13.56

and
Host 67, (an Intel P4 1.8, northwood, 512M ram, Winxpsp2) has done:
1484 1449 16 Feb 2006 0:38:05 UTC 16 Feb 2006 20:59:37 UTC Over Success Done 3,668.64 5.32 5.32
1475 1440 16 Feb 2006 0:38:05 UTC 17 Feb 2006 5:10:13 UTC Over Success Done 3,517.48 5.10 5.10
1474 1439 16 Feb 2006 0:38:05 UTC 17 Feb 2006 5:10:13 UTC Over Success Done 3,503.78 5.08 5.08
1473 1438 16 Feb 2006 0:38:05 UTC 17 Feb 2006 3:26:39 UTC Over Success Done 3,729.38 5.41 5.41
1472 1437 16 Feb 2006 0:38:05 UTC 17 Feb 2006 3:26:39 UTC Over Success Done 3,562.56 5.17 5.17
1471 1436 16 Feb 2006 0:38:05 UTC 17 Feb 2006 1:10:55 UTC Over Success Done 3,874.70 5.62 5.62
1470 1435 16 Feb 2006 0:38:05 UTC 17 Feb 2006 0:06:03 UTC Over Success Done 3,278.38 4.76 4.76
1469 1434 16 Feb 2006 0:38:05 UTC 17 Feb 2006 0:06:03 UTC Over Success Done 3,540.36 5.14 5.14
1468 1433 16 Feb 2006 0:38:05 UTC 17 Feb 2006 0:06:03 UTC Over Success Done 3,412.34 4.95 4.95
1465 1430 16 Feb 2006 0:38:05 UTC 16 Feb 2006 20:06:49 UTC Over Success Done 3,594.17 5.21 5.21
1464 1429 16 Feb 2006 0:38:05 UTC 16 Feb 2006 18:45:39 UTC Over Success Done 3,697.19 5.36 5.36

I will take the time to analyse this in the next post.

tony

First off, I see that not one WU has errored out for any reason.

Secondly, I see 3 of my 4 hosts are taking about an hour for each WU, but strangely, my Celeron is all over the map. The celeron is controlled via VNC so I don't do anything with it at all except let it sit under a table and crunch away. I check it about once a day to report results, and make sure it still running.

[edit] I have changed my Cpu run time to 2 hours and will continue to run a while. I'm off to update all hosts.

tony

Mr, Aaron finney, I see what you mean.

I updated the Ralph prefs and confirmed that "my account" showed my change to 2 hours. Then I went to each host, and updated the "Ralph" project from the "projects tab" of the Boinc manager. In each case I saw NO mention of an updated pref on my messages tab.

[edit] I've waited 20 minutes and still don't see it(on any host), even on hosts which are "requesting work". Something's wrong. Even if manual update didn't work, it certainly should have been seen/added to the message log when requesting work. I'll keep an eye on future WUs to see if it was accepted by the host client.

Have I understood this discussion correctly? If I download WUs as before they will run a prediction and then end after a varaiable amount of time and return the results.

they'll end after they finish their current "answer" (unless they predict that they'll go over the time limit by a long way, and they'll stop before however many hours you've set)

If I set the CPU run time preference it will run sufficient predictions on the same WU to process up to that amount of time (Einstein runs different calculations on the same data, but does not have a time limit)

correct :)

If so what happens to the partial result? I.e. if I set it to 2 hours then I assume the work that is cut off at 2 hours will only be a partial result. Is this prediction ignored?

not quite, a normal WU would do 10 predictions (10 "answers"), the new method will keep working and producing "answers" untill the time limit is reached/exceeded
however, if it knows that each answer takes say 20 minutes (just a guess), and you've set a limit of 2 hours, if it reaches 1:50 then it might stop there, and not do another one (because it would run over the limit quite a bit)

I.e. should we all set such a CPU limit?

use it how you would use rosetta, set your prefs to match your needs
this way a veriety of settings get tested (i'm sticking with "none selected" because i don't have bandwidth issues, and to make sure that the default works as desired for those who don't choose a pref)
if you need to use 8 hours, use that, if you need 2, use that, if it doens't matter, use "none seleceted"
i'm sure the devs will work things out as results are returned (but it's looking good so far)

Well, even if the message board never displayed a change in prefs, my host #66 is now taking two hours to do a wu. I even rechecked the log and it's NOT there. At no time from when I change the website until it actually crunched the WU did the messages tab display that an updated pref was found.

Well, even if the message board never displayed a change in prefs, my host #66 is now taking two hours to do a wu. I even rechecked the log and it's NOT there. At no time from when I change the website until it actually crunched the WU did the messages tab display that an updated pref was found.

I've found the same for other projects too, unless it's something major that's obvious (like resource share) there's no message about it, but BOINC is aware of the change (looking at the TCP stream of the GUI RPC) so i'm guessing this is a boinc client issue rather than a project issue

I've found the same for other projects too, unless it's something major that's obvious (like resource share) there's no message about it, but BOINC is aware of the change (looking at the TCP stream of the GUI RPC) so i'm guessing this is a boinc client issue rather than a project issue

Hmmm, not seen this mentioned at any of the meetings. Maybe I should whip up an email and inquire.

[edit]I placed an email on Boinc Dev about this.

Aaron,
This is probably a wrong assumption as well, but have you set a particular venue for your machine and did you then update the preferences for that venue. Obviously if you are using only the default preferences then you would have made the change correctly and it should have shown up in the messages tab. But if you have your computer set as lets say HOME for the venue and then update only the default preferences, when you update, the preferences change may not show up in the messages tab becaue they would not apply to that venue.

No, this isn't what I'm talking about either, it's the work unit run time preference setting only, and all machines have the same venue for that preference.

In any case there may be a problem with this area anyway. The setting that controls Work Unit run time is unique to Ralph and may not be properly recognized by BOINC as a preference change. This would be a bug, but I do not know if it could be fixed from the project side or not. There should be a way to trigger this from the Ralph application however.

This is the correct scenario, and you're right, I don't know how to trigger BOINC to issue a message like that either. Could be a client side bug, or even server side.

---

This is the correct scenario, and you're right, I don't know how to trigger BOINC to issue a message like that either. Could be a client side bug, or even server side.

it's just a case of the client not displaying a message about it, because as i stated, the client is actually aware of the change, it just doesn't make any noise about it

We are going to update the code to remove the 99 limit. For most work units 99 would take too long. We would prefer lower run times so that results are returned quicker.

Has that been done? Because I'm crunching some workunits right now that are up to Model # 147.

2/17/2006 11:56:59 PM|ralph@home|Starting result BARCODE_30_1a19A_209_15_0 using rosetta_beta version 484

That one.

---

The Remaining CPU time is not consistant with the amount of time that SHOULD be left.

I.E. A workunit that has processed for 14:30:00 SHOULD have 1:30:00 left before the end of it's processing, at 16:00:00, with a preference set of 16 hours.

The time instead is reporting 2:06:00 remaining, which is 0:36:00 too much. I'm assuming that it is estimating the amount of time it will take to finish the particular model that it is on, and that is where the extra 36 minutes comes from. If that is the case, then I guess it's working as intended, but shouldn't it stop on the model that keeps it UNDER the amount specified in the preferences? IMHO, if the preference is 16 hours, it shouldn't go over 16 hours. Of course, you could always just re-word the preference to read "Approximate CPU Run time Preference", and then you can appease both sides of the fence without making any programming changes.

---

OK I got an error now. My Celeron 500, 256M ram, Win98SE, Ralph Host 103 has had a computation error. This unit produced 20 successful returns with the Ralph time pref set to "not selected", the first WU with it set to 2 hours has had a Computation error at 02:01:00. this is the WU

5133 4850 18 Feb 2006 13:42:21 UTC 19 Feb 2006 0:02:27 UTC Over Client error Computing 7,260.00 4.85 ---

I suspect it has to do with the way win98se uses a wall clock.

tony

NOTE: Further up/down in this thread is more on this hosts set up and history

Edit, the first 20 were 4.83 and the ones of this run are 4.85, don't know if it matters, but there it is.

Also, the one that failed was not on my machine when I updated prefs, so it wasn't a matter of changed prefs while running.

Note: I changed my prefs back to "not selected" on CPU run time. did the project update, still don't see it on the messages tab, then suspended the only other project on host 103, Ralph is now running. I'll know in an hour whether or not it's a 4.85/4.83 thing causing this computation error.

As I sit here staring at the screen, I wonder:

The other difference in WUs on host 103 are that they were 4.83 HBLR wus that ran under my "not selected" preference, and the new ones are 4.85 Barcode WUs. This first WU after I changed back to "not selected" sat at 1% for 16 minutes, then sat at 1.78% for 10 or so minutes, then the percentage quickly ran to 51.67% were it stay for nearly 20 minutes and just now (at the 1:00:00 mark literally jumped to 74%.

I'm coming to the conclusion that these wus are different and I can't really use any comparisons between 4.83 successes and 4.85 failures to find any meaningful conclusions (read I'm wasting my time trying). I need to compare apples to apples and "Barcodes" to "Barcodes". Am I right here?

Another Question:

When my scheduler requests work, are the requests based upon the existing CPU time preference setting? Is changing that Preference setting after work is on your puter supposed to work on existing WUs? Or will we have to wait to burn through the existing wus before our preferences work again?