[QUOTE=chalsall;461452]I'm not quite sure what you are seeing, but that report updates in "real-time", and generally changes several times per hour.

Could you possibly instead mean [URL="https://www.gpu72.com/reports/available/p-1/"]this report[/URL]? If so, then you are correct; it rarely changes.

This is because the TF'ing is ahead of the P-1'ing, so the only time candidates of that class appear is when we enter a new range, and/or someone does an "independent" P-1 run.[/QUOTE]

No, the few numbers I am watching are the 83M & 84M to 76-bits. Both have been steady at 999 & 1000 respectively with nothing going to P-1. I have fed the 84M to 76-bits (slowly) and nothing seems to change. Perhaps you have some throttling in place that maintains a constant value at 76-bits.

[QUOTE=RichD;461525]No, the few numbers I am watching are the 83M & 84M to 76-bits. Both have been steady at 999 & 1000 respectively with nothing going to P-1. I have fed the 84M to 76-bits (slowly) and nothing seems to change. Perhaps you have some throttling in place that maintains a constant value at 76-bits.[/QUOTE]

Oh, I understand. And yeah, you are correct.

The GPU72 P-1'ers are currently working on the 81M and 82M ranges; we should finish 81M in about two weeks, and 82M about a month later.

I could "throw back" the 999 candidates currently reserved in 83M, but I have them there in case anyone would like to explicitly ask for P-1 candidates already TF'ed to 76 bits.

Lastly, yes, my "unreserve" script keeps 1,000 candidates in 84M TF'ed to 76 bits, and returns the rest.

The SSL cert I'm seeing for [url]https://www.gpu72.com/[/url] expired a few hours ago, interrupting communication.

[QUOTE=Mini-Geek;463920]The SSL cert I'm seeing for [url]https://www.gpu72.com/[/url] expired a few hours ago, interrupting communication.[/QUOTE]

Hey guys. Sorry. This was supposed to be automatic. It wasn't...

A great way to spend a Friday night. Sometimes I think computers are sentient.

[QUOTE=chalsall;463943]A great way to spend a Friday night. Sometimes I think computers are sentient.[/QUOTE]

Hey. So GPU72.com's SSL's connections should be working again. Please let me know if anyone sees any issues.

I trusted the "Let's Encrypt" script to work, but it didn't (probably my fault, of course). Caused a bit of an issue for a scheduled outing in "meat space".

Just putting this out there... I truly believe that computers are sentient, and know when you are in front of an audience. Then they giggle while they refuse to work correctly as they always did while you tested them before.

LOL...

Sentient and Malign! aka S&M.

The following completed yesterday, 2017-08-03, and shows this on PrimeNet. However, it still shows as an assignment on GPU72. The DC was acquired though P95, but ran on a GPU and was turned in manually.
[URL="https://www.mersenne.org/report_exponent/default.php?exp_lo=40988699&full=1"]40988699[/URL]

[QUOTE=kladner;464836]The following completed yesterday, 2017-08-03, and shows this on PrimeNet. However, it still shows as an assignment on GPU72. The DC was acquired though P95, but ran on a GPU and was turned in manually.
[URL="https://www.mersenne.org/report_exponent/default.php?exp_lo=40988699&full=1"]40988699[/URL][/QUOTE]
The number in question has now crossed over to completed status. :smile:
EDIT: Kind of like crossing the river Styx?

[QUOTE=kladner;464874]The number in question has now crossed over to completed status. :smile:[/QUOTE]

OK, I haven't drilled down, but LL/DC candidates acquired through the proxy but are then submitted manually take longer to detect.

Thanks, Chris.
I have two more pending, now. One very recent, ~3:50 PM UTC: 40997323
41003399 is maybe 30 hrs since submitted.
Given the talk about how far ahead factoring is, I moved both GPUs to DC work.
Performance:[INDENT]i7 6700K @ 4.3 GHz, 4 cores, one worker: ~2.2 ms/it, depending on other tasks running. 43.8M exp.
GTX 1060 @ ~2.0 GHz, ~4.16 ms/it, depending on ambient temp. 43.8M exp.
GTX 460 @ 855MHz(!), 8.36ms/it, little variation, 43.8M exp.[/INDENT]

[QUOTE=kladner;464953]I have two more pending, now. One very recent, ~3:50 PM UTC: 40997323. 41003399 is maybe 30 hrs since submitted.[/QUOTE]

OK, I drilled down on the code, and increased the number of assignments which are checked for completion. Such candidates should now have the work detected within a couple of hours.

[QUOTE=kladner;464953]Given the talk about how far ahead factoring is, I moved both GPUs to DC work.[/QUOTE]

Coolness! Thanks.