Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
![]() |
World Community Grid Forums
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
No member browsing this thread |
Thread Status: Active Total posts in this thread: 38
|
![]() |
Author |
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Yeah, right, "The "Too Many Errors" issue is resolved"", but only in my dreams. At about 12:45UTC hit the Refresh Top ResultsXML update button and poof the ValidationState=3 records were gone, no trace on the Result Status page, no way of knowing whether the friends of Mr.Cairo have done what they were supposed to do... give credit before shooting them down the rabbit's favorite hole, not to speak of not being in the current extraction set, nothing to sent to the archive db.
(As there are automaton donkeys to count with, had prudently made a backup, and retrieved the records there, so maybe tomorrow morning can see if these constitute a / the difference [Maybe these were SgtJoe's difference he posted about instead of the nightly spill, but do not know if he's doing Android crunching] |
||
|
Sgt.Joe
Ace Cruncher USA Joined: Jul 4, 2006 Post Count: 7697 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
[Maybe these were SgtJoe's difference he posted about instead of the nightly spill, but do not know if he's doing Android crunching] The "spill" was my difference so that issue is fully resolved. I am not crunching on android as when it comes to phones I somewhat of a Luddite. I have a mobile phone, but it only does one one thing - allows phone calls. ![]() Cheers
Sgt. Joe
*Minnesota Crunchers* |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Too stupid of course to sit there and watch it, so the [partial] solution is a type of cronning:
----------------------------------------Automatically refresh data at regular intervals Click a cell in the external data range. Click Data > Connections, click the arrow next to Refresh All, and then click Connection Properties. Connection Properties Click the Usage tab. Select the Refresh every check box, and then enter the number of minutes between each refresh operation. So it's set to update every 720 minutes and then run the archiving routines [haha first run has to be synched to UTC 12:06] but improvement on that is underway; Run at certain times: http://www.globaliconnect.com/excel/index.php...p;catid=79&Itemid=475 BTW Did -NOT- get credit for the Too Many Errors that disappeared yesterday noon-ish [Valid but no non-error wingman to arrive in time]. This morning yet another 8.91 hours are lurking front the (A)Rabbiata hole. Just in case, when I am visiting, now a pink button can be pushed :\ New pic of v1.4a... major blanding down on the colors and squares-saw puzzling to fit it best and logical. http://i137.photobucket.com/albums/q210/Seker...utionApp1.40.png~original Guess which is the most important 'square' of those? :0)) Edit: And dash-boarded, to make the square-fitting free-form rather than being locked into compromising the column width grid. [Edit 1 times, last edit by Former Member at Jul 20, 2015 12:00:35 PM] |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Closing this project off, recon auto-perfect [And still molto arrabbiata over re-discovering there's no crediting for time or results on the Android OET1 tasks which are computing perfectly on my tablet, but encounter 5 others with error or no reply]. Consistency of policy and practice is non-existent [Still].
The learneth "Oh yes, that's how it was" is now being backported [too hot to be outside, or risk heat stroke]. Very old code used for the Performance Charts is redone. You find that in yonder days when starting with VBA, you just record actions and if -Always- reproducing the desired outcome with Alt-F8 > Select > Run, you move on to the next bit to automate, no beautification. You get to see the difference of how better looking and readable it becomes, but mostly the efficient and speedy code execution [moving sheets and cells is process costly]. An example Before: ' 7 Day centered formula for CEP2 MacroAfter: ' 7 Day centered formula for CEP2 Macro This executes without even having time to blink an eye, it's background stuff anyway, so who cares ![]() Till next. |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
The closing off took a bit more as a
![]() Rebuild the way history is contructed... rather than inserting row near top of table, verify formula are fine [Excel pulls rabbits from hats too and will happily populate them from some hidden original version], checking formula and navigating around pulling next project and next project, making it slow, just take a copy and drop the values one record down [which freezes the last record for good at that ridding of ever increasing number of index-match formula]. The beauty of this with tables being auto-extended when immediately below or right new data appears, no worries about ranges [they are with tables also auto-named per the headers]. Then, since reusing ancient code from the previous decade, found ways to half the navigation in input [costly in time factor] and again to half of that et voilá, the whole done in not under 30 seconds, not under 22, now done in 14.6 seconds. ![]() Original code [Beetle]: Range("ActiveQry").Select ' Select ActiveQry, *first line of data*.Improved code [F1] Range("ActiveQry").Select ' Select ActiveQry, *first line of data*.Ultimate code [F1-2017] For Each prj In Range("ActiveQry") From 20 lines, to 15 lines to 10 lines, the only feedback while running in the status bar informing on which history is being extended. Of course, exposé, the constructive criticasters can improve on this, but over at StackOverflow indications were marginal, of course, the DB is still small, but then this only adds 104 records per annum [one every 12 hours]. Arraying the Range Object would probably just take fractions off [arraying avoids accessing data repeated, transacting in memory and only when done, write the result back. ![]() v1.5c being archived. ![]() |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Came out, the nightly spill times are not 00:06 and 12:06, but close to 00:06:09 and and 12:06:09 [who invented that :?]. Had one CEP2 job report at 00:05:55 and validate at 00:06:03 and out popped a difference of 11:46 hours against the MC page. Hunting model duly modded.
![]() Latest v1.5c image , with same micro-graphs added. Like it says at BOINCStats too, the lower the better [if you fancy the rankings] :D) Think I'm n ow truly done with this. ![]() |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
No, No... No, No, No, it's 10 seconds more, not 9, and the bafflement is when thinking to be really sharp and pick up the actual timestamp from the community pages to compute the 'real' spilltime... it said:
----------------------------------------Statistics Last Updated: 7/29/15 12:06:02 (UTC) [2 hour(s) ago] Had a task validating at 12:06:10, but it was nevertheless included in the morning results total. ![]() Looks like the procedure order is possibly sent to a multithreaded queue and then individual processes cutting off sooner or later, but not in perfect order i.e. someone tells to take a closing timestamp, when the 4th in the relay has not crossed the line yet with the baton. [As what the techs on the science apps would refer to as a 'Race Condition' ![]() Basically, there is for moi no programming against this, just a permanent case of throwing bloodied towels at so much consistent inconsistency [who cares anyway ![]() [Edit 1 times, last edit by Former Member at Jul 29, 2015 3:36:58 PM] |
||
|
Sgt.Joe
Ace Cruncher USA Joined: Jul 4, 2006 Post Count: 7697 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
[who cares anyway ![]() Some of us do care and we appreciate your work on this. Cheers
Sgt. Joe
*Minnesota Crunchers* |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Since, I've developed the "Recon Nook", where all tasks are listed that have a validation state = 1 and a modtime from 12:00:00 to 12:06:10 [same for the midnight block], with the totals incrementing in a second column, the total that matches the MC History page day record is the true cut-off point. I've got only few, but those that have dozens of devices will have a hard time to find which one is in and which one is out. If having dozens, you are likely only interested in ballpark anyways.
Regrettably, since there is no XML format for the history where you can specify the verification code and member name in the query, it's having to go on the website and do the old scrape to get the day record... the query now stops at the password field. So far not found a way to open the pw field, it's grayed out in the xml query connection screen on Excel, to include the tick-box 'Remember password'. I'd probably have to get into odc coding to get around that, and to be honest, not really interested iiat this time to again learn a scripting format. Oh wait, Office uses IE as default... setting up the query a IE page showed up, with the WCG page to tick the 'Remember' box, the IE asked if I wanted the PW to be remembered or not. Did so. Exiting and launching the app, hit refresh and no PW question. Now that cookie sits there untill the cleaner comes along for it's weekly wipe. Can't win them all, and when 'The List' item on adding the sort var for modtime to the API comes to getting a 'Done' tickmark... next decade maybe [at least, jhindo gave the impression WCG would still be there]. Right now the data arrives in received time order i.e. the spill list is not in time of change order... one more thing to fix locally I suppose. |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Something IS working... just now created a Beta tab in the hunting book. As having developed two templates, single app, which most sciences are, and the 2 app FA@H which I suspect has the currently tested BEDAM tested science fold in [and hoping the techs will give it in that case a moniker that also starts with FA], used that one in copy/insert and was done in under a minute. As the Beta apps are sequential, a simple substitute of the FA* filter key with BETA* will pick up all now and future tests.
----------------------------------------Oh, the MC dashboard now has a true graph included for runtime (still in seconds) and results. Later has a downward trend and today plummeted with the beta monsters running on 14 of 18 cores [2 only get OET, se tablet, and 2 reserved for CEP2]. The Recon Nook worked perfectly lifting out a 1 result spill again. Just hit the archive button and moved onto afternoon stats accumulation. Next trick to develop is scripting a consolidation, as now certain checkfields on global don't know yet of Beta... something along the lines of "add all tabs that do not start with All in the name."... has to be else there'd sneak in a circular and iteration would increase values endlessly by the power 2 ![]() Carry on. ![]() (No screenshot... trust me, it's real ![]() [Edit 1 times, last edit by Former Member at Aug 3, 2015 2:08:07 PM] |
||
|
|
![]() |