Still toying with the tool, it's grown out of proportion to be honest, but it's fun to find the limits, and move them. Latest addition, an error analysis [which in my next dream will hotlink to the actual result status record for detail... click and hypervisor over]. Todays rough out...



The timestamp is WCG's converted to local time, so it may correlate to something daft being done at the time of the task's demise. Host names hidden ;>)

For me, just mosey over to the Error Report tab, but put your coffee down first and keep the prosac ready [pre-emptive strikers take one first (someone said crunching was fun .]

(And Marshmallow installed on the tablet last Friday and since, zero, not, null, nada errors, even while crunching on battery while gaming.)

One better, show result name and on hover see the record, click and jump to the full detail... muahahaha

Of course, when maxing out doing my own stuff, BOINC gets to play second fiddle....

3016 11/12/2015 5:55:08 PM Suspending computation - CPU is busy
3017 World Community Grid 11/12/2015 5:55:08 PM [cpu_sched] Preempting E234668_908_S.216.C27H31N1S1.BPMDLFDZFGSEEZ-UHFFFAOYSA-N.18_s1_14_0 (left in memory)
3018 World Community Grid 11/12/2015 5:55:08 PM [cpu_sched] Preempting ugm1_ugm1_19800_0397_0 (left in memory)
3019 World Community Grid 11/12/2015 5:55:08 PM [cpu_sched] Preempting ugm1_ugm1_19800_0970_1 (left in memory)
3020 World Community Grid 11/12/2015 5:55:08 PM [cpu_sched] Preempting ugm1_ugm1_19800_1959_1 (left in memory)
3021 World Community Grid 11/12/2015 5:55:08 PM [cpu_sched] Preempting ugm1_ugm1_19800_2136_1 (left in memory)
3022 World Community Grid 11/12/2015 5:55:08 PM [cpu_sched] Preempting ugm1_ugm1_19800_2105_0 (left in memory)
3023 World Community Grid 11/12/2015 5:55:08 PM [cpu_sched] Preempting ugm1_ugm1_19804_0641_0 (left in memory)
3024 World Community Grid 11/12/2015 5:55:08 PM [cpu_sched] Preempting ugm1_ugm1_19808_1947_1 (left in memory)
3025 11/12/2015 5:55:18 PM Resuming computation


At least, the tasks don't get unloaded, nor crash due time-outs, which CEP2 is rather sensitive to.

'Over the top' it's gone for sure, but still fun... if there's a blockage on progress, I will find a solution. Today "On Error Resume Next". Sometimes a scraped page is not reachable, so the script hangs [yes we still have stuff needing scraping the old previous century way]. This one just skips the first offending [non-critical] line, continuing on... stats ran very long last night :(.

Most irritating was that in Excel the result archiving fetches a filtered record set [and we're limited by WCG to 250], but if none of the records meet the conditions, Excel copies the whole filtered out set instead of copying nothing [skip on or present an error code to handle]. An additional [initial ugly code] test determines if filtered map results in zero, then continues with the next page [max 250 again], in the loop.

Old days '>>', append additional input to existing text file, circa 1990 it must have been I last used that function, but nothing wrong with it :O)

Crisis what crisis? The testing data set expanded from < 600 results to 9000 overnight and Office did not like that [the way WCG offers a max of 250 per Results request]. Fortunately I knew where to look and hit Google 100% score on the top item, 5 minutes understanding, 30 minutes preparation of an input map into a for-next, discovering how many records to come [the API tells you on the first record received], and off to listobject sizing, this time the testing was only 5 minutes of the ~60 total time spent.

Only now noticed MSDN-Dev Center is using GitHub for development and documenting, same as Berkeley BOINC is kept on]. The development speedup actually came mostly from a code I wrote yesterday, automating the column sizing for 34 fields, which runs in less than a blink of an eye [A consistent presentation of the data view is always pleasing ;P]. Re-usable code is great... another module added to the Hunting too, v2.0o

Happy birthday WCG, back to a run in 2 minutes, over copper telephone wire.