We used to have fair morning-afternoon symmetry. Not looking at the full day when fetching the number, it does look odd as then it implies a carry over from previous day [unlikely], or a late morning stats run, but adding up it doesn't.

As for the points, ignore the present... they're substantially inflated presently as I posted few days ago... new apps for zika+hst1, the cycle usually starting off hi and then coming down.

Something is cooked anyway, as several FAH2 which are not wingman dependent are still in PVal.

FAH2_ 000090_ avx38781_ 000010_ 0015_ 030_ 1-- 3113135 Pending Validation 7/9/16 19:41:08 7/10/16 21:20:30 13.86 / 13.99 214.7 / 0.0
FAH2_ 000094_ avx38785_ 000093_ 0014_ 027_ 0-- 3113135 Pending Validation 7/9/16 19:36:31 7/10/16 19:33:01 13.64 / 13.83 212.3 / 0.0

It's 22:07 UTC now... other sciences seem to have up-to-date validations though (maybe all those -186 errors (aka -119/-120) caused the device to be considered not trustworthy?

I managed to report results by setting "no new tasks" and adding
<max_tasks_reported>1</max_tasks_reported>
in cc_config, as Sekerob suggested.

When all work was reported I set work buffer to 0.01 days and CPU% to only use 1 core. This setting downloaded just 1 WU. Then increasing CPU% by 1 core at a time I downloaded more WUs. Now everything is back to normal.

To quote you Sekerob: "Any setting that ensures the smallest possible work requesting."
Forcing Boinc to only use 1 core did the trick!