Is calculation of hours run correct?

log in

Advanced search

Message boards : Number crunching : Is calculation of hours run correct?

Previous · 1 · 2 · 3
Author Message
Dr Who Fan
     
Avatar
Send message
Joined: 29 Jul 11
Posts: 253
Credit: 970,913
RAC: 282
Total hours: 1,350,806
Message 9063 - Posted: 24 May 2022, 8:20:42 UTC - in response to Message 9061.

Appears the Home Page stats for "last day hours" has been fixed.
____________

mikey
   
Avatar
Send message
Joined: 20 May 10
Posts: 341
Credit: 1,220,168
RAC: 776
Total hours: 5,018,505
Message 9064 - Posted: 24 May 2022, 12:27:57 UTC - in response to Message 9063.

Appears the Home Page stats for "last day hours" has been fixed.


Mine too!!

Dr Who Fan
     
Avatar
Send message
Joined: 29 Jul 11
Posts: 253
Credit: 970,913
RAC: 282
Total hours: 1,350,806
Message 9135 - Posted: 8 Jun 2022, 15:23:08 UTC - in response to Message 9053.

Seeing this again....

My home page stats hours https://wuprop.boinc-af.org/home.php are still showing task hours from several days ago counting as "last day hours", but the "last 24 hours hosts" hours page are correct.

____________

Dr Who Fan
     
Avatar
Send message
Joined: 29 Jul 11
Posts: 253
Credit: 970,913
RAC: 282
Total hours: 1,350,806
Message 9136 - Posted: 9 Jun 2022, 15:35:21 UTC - in response to Message 9063.

Appears the Home Page stats for "last day hours" has been fixed.

Still showing hours of tasks that have not been run in days on my home page... example:
PRIVATE GFN SERVER GFN-11 MEGA Prime Search 10,094.92 0.07

____________

mikey
   
Avatar
Send message
Joined: 20 May 10
Posts: 341
Credit: 1,220,168
RAC: 776
Total hours: 5,018,505
Message 9137 - Posted: 9 Jun 2022, 22:39:54 UTC - in response to Message 9136.

Appears the Home Page stats for "last day hours" has been fixed.

Still showing hours of tasks that have not been run in days on my home page... example:
PRIVATE GFN SERVER GFN-11 MEGA Prime Search 10,094.92 0.07


I too get it for the Private GFN Project plus for Latin Squares ODLK3@home:

PRIVATE GFN SERVER GFN-11 MEGA Prime Search 1,599.42 15.65

latinsquares odlk3@home 110,670.28 0.25

mikey

Dr Who Fan
     
Avatar
Send message
Joined: 29 Jul 11
Posts: 253
Credit: 970,913
RAC: 282
Total hours: 1,350,806
Message 9157 - Posted: 24 Jun 2022, 2:16:38 UTC

Not sure what is causing NEGATIVE HOURS on some sub-projects according to the Last 24 hours user activity page?

*Could it be Database / delayed hours counting problems by server code?
Here is one example from my page loaded a few minutes ago:


"HOST 5"
LODA / LODA -10.95
NFS / 16e Lattice Sieve for smaller numbers -2.32
Ramanujan Machine / Ramanujan Machine -0.92

____________

Profile Henk Haneveld
   
Send message
Joined: 14 Nov 17
Posts: 5
Credit: 174,660
RAC: 114
Total hours: 373,635
Message 9162 - Posted: 25 Jun 2022, 13:46:00 UTC - in response to Message 9136.

Appears the Home Page stats for "last day hours" has been fixed.

Still showing hours of tasks that have not been run in days on my home page... example:
PRIVATE GFN SERVER GFN-11 MEGA Prime Search 10,094.92 0.07

Is still NOT fixed.

Link
 
Avatar
Send message
Joined: 20 Jun 12
Posts: 57
Credit: 293,334
RAC: 70
Total hours: 145,594
Message 9164 - Posted: 25 Jun 2022, 15:53:34 UTC - in response to Message 9162.

Is still NOT fixed.

There are lots of things not fixed or actually breaking down more and more every day as it seems.
____________
.

mikey
   
Avatar
Send message
Joined: 20 May 10
Posts: 341
Credit: 1,220,168
RAC: 776
Total hours: 5,018,505
Message 9168 - Posted: 26 Jun 2022, 10:05:34 UTC - in response to Message 9157.
Last modified: 26 Jun 2022, 10:31:01 UTC

Not sure what is causing NEGATIVE HOURS on some sub-projects according to the Last 24 hours user activity page?

*Could it be Database / delayed hours counting problems by server code?
Here is one example from my page loaded a few minutes ago:

"HOST 5"
LODA / LODA -10.95
NFS / 16e Lattice Sieve for smaller numbers -2.32
Ramanujan Machine / Ramanujan Machine -0.92


ALL of mine are NEGATIVE hours this morning!!!

33 SiDock@home CurieMarieDock on BOINC + zipped input, checkpoints and progress bar 45,499.95 -114.62
37 Cosmology@Home camb_boinc2docker (beta test) 37,225.65 -24.00
26 Radioactive@Home Radioactivity Monitor 50,567.63 -16.92
52 SRBase TF 25,047.85 -9.95
97 iThena.Measurements iThena OONI Probe 7,903.12 -3.98
88 SRBase Sierpinski Base 10,161.13 -3.67
63 iThena.Measurements iThena PERF 13,296.82 -3.40
19 Milkyway@Home MilkyWay@Home 54,680.93 -3.00
96 climateprediction.net UK Met Office HadAM4 at N216 resolution 7,944.12 -3.00
35 Einstein@Home Gamma-ray pulsar binary search #1 on GPUs 38,932.60 -2.97
187 RNA World cmsearch VM (VirtualBox) 1.0.2 332.45 -1.00
18 latinsquares odlkmax@home 61,867.08 -0.82

And now 10 minutes later all is fine again!!

33 SiDock@home CurieMarieDock on BOINC + zipped input, checkpoints and progress bar 45,544.30 2,568.97
37 Cosmology@Home camb_boinc2docker (beta test) 37,225.65 555.45
26 Radioactive@Home Radioactivity Monitor 50,573.62 377.05
35 Einstein@Home Gamma-ray pulsar binary search #1 on GPUs 38,934.60 168.22
155 Einstein@Home Binary Radio Pulsar Search (Arecibo, GPU) 2,018.33 117.58
97 iThena.Measurements iThena OONI Probe 7,905.12 85.50
52 SRBase TF 25,047.85 79.18
96 climateprediction.net UK Met Office HadAM4 at N216 resolution 7,947.12 68.85
63 iThena.Measurements iThena PERF 13,296.88 64.77
19 Milkyway@Home MilkyWay@Home 54,680.93 51.95
88 SRBase Sierpinski Base 10,161.13 27.37
187 RNA World cmsearch VM (VirtualBox) 1.0.2 333.45 22.95
18 latinsquares odlkmax@home 61,867.08 18.53
177 GPUGRID Python apps for GPU hosts 864.35 15.33

Dr Who Fan
     
Avatar
Send message
Joined: 29 Jul 11
Posts: 253
Credit: 970,913
RAC: 282
Total hours: 1,350,806
Message 9191 - Posted: 1 Jul 2022, 1:56:41 UTC

Found MORE problems with hours calculations:
All my PC's and Android devices are missing total hours on average 2 hours per core/thread short per 24 hour wall time.

These two apps are just easy to discover examples. Each app has been running 24 hours wall time on 2 cores each for a total of what should be 48 hours each.

96 PrimeGrid Sierpinski/Riesel Base 5 Problem (LLR) 4,076.47 44.00 125 PrimeGrid Cullen (LLR) 2,572.80 44.00


Instead, as you can see above, they only received credit for 44 hours each.
That's 4 hours each I did not receive credit for.
Looking at WuProp completed tasks run times, upload and download times I can see some task run less than 3600 seconds, sometimes more, and then there's the upload/download time delays consuming a few seconds each.

Maybe it's time we go back to 6 hour Runtimes on tasks so that the loss of calculated hours is minimized to a few minutes per calculated 24 hour wall time.
____________

Dr Who Fan
     
Avatar
Send message
Joined: 29 Jul 11
Posts: 253
Credit: 970,913
RAC: 282
Total hours: 1,350,806
Message 9215 - Posted: 8 Jul 2022, 21:37:32 UTC - in response to Message 9191.

Problem is still ongoing... ... Any chance to fix?

PrimeGrid Sierpinski/Riesel Base 5 Problem (LLR). 4,374.27 44.00
PrimeGrid Genefer 19 4,313.83 44.00

Above two examples are running on 2 threads each and should be registering 48 cumulative computational hours. Each one missing 4 hours over 24 hours wall time reporting.

Didn’t have the problem under previously run 6 hours tasks, only since project resumed under 1 hour tasks.
____________

Profile [AF>WildWildWest]Sebastien
     
Dictator
Avatar
Send message
Joined: 28 Mar 10
Posts: 2212
Credit: 474,332
RAC: 51
Total hours: 1,169,640
Message 9217 - Posted: 9 Jul 2022, 8:58:49 UTC - in response to Message 9215.

There is nothing to fix. Counting hours runs correctly.
Check your host activity, you'll see counting hours is OK.

Last day count is an approximate computing.
____________

mmonnin
     
Send message
Joined: 22 Aug 16
Posts: 372
Credit: 1,650,050
RAC: 650
Total hours: 7,323,354
Message 9218 - Posted: 9 Jul 2022, 11:33:06 UTC

Using <report_results_immediately>1</report_results_immediately> would reduce the time between tasks
https://boinc.berkeley.edu/wiki/Client_configuration

Dr Who Fan
     
Avatar
Send message
Joined: 29 Jul 11
Posts: 253
Credit: 970,913
RAC: 282
Total hours: 1,350,806
Message 9219 - Posted: 9 Jul 2022, 13:07:13 UTC - in response to Message 9218.
Last modified: 9 Jul 2022, 13:18:15 UTC

Using <report_results_immediately>1</report_results_immediately> would reduce the time between tasks
https://boinc.berkeley.edu/wiki/Client_configuration

THAT IS ALREADY IN USE across all devices so it's something else.

Looking at Run Time of reported task most run 3600 seconds, others run a little as aprox 3200 seconds and some more than 3600 seconds... if less they are being upload/download at aprox 3600 seconds.
Example: Task 124639738
Sent 9 Jul 2022, 9:21:44 UTC
Received 9 Jul 2022, 10:22:22 UTC
Run time 54 min 31 sec
Application version Data collect version 4 v4.29 (nci)
____________

mikey
   
Avatar
Send message
Joined: 20 May 10
Posts: 341
Credit: 1,220,168
RAC: 776
Total hours: 5,018,505
Message 9223 - Posted: 10 Jul 2022, 23:11:21 UTC - in response to Message 9217.

There is nothing to fix. Counting hours runs correctly.
Check your host activity, you'll see counting hours is OK.

Last day count is an approximate computing.


So if I copy and paste the total numbers for a particular application into a spreadsheet every day am I'm getting the most accurate numbers possible at the moment of the webpage being generated? And is it as accurate as possible day after after day?

Previous · 1 · 2 · 3
Post to thread

Message boards : Number crunching : Is calculation of hours run correct?


Home | My Account | Message Boards | Results


Copyright © 2022 Sebastien