lowest 10M exponents available - pick your choice

[ckdo]

http://v5www.mersenne.org/report_ll/?exp_lo=33600000&exp_hi=33605000 lists 33600571, 33601693, 33602077, 33602111 and 33602141. I'd call that "checked in".

[Brian-E]

Originally Posted by tallguy

Every time I check it gets pushed out a bit more since the wife uses it on battery so much. Best guess now is April.... 2023.

In all seriousness, mid-April.

One thing which you may have already noticed is that the currently stable versions of Prime95/mprime tend to give estimated completion times which are too optimistic if you don't have the computer running all the time. The reason for this is that the RollingAverage term in the local.ini file is not correctly calculated: if Prime95/mprime is running only a limited number of hours per day, RollingAverage will set itself too high. This is the four digit number which is a correction value for CPUhours with 1000 meaning no adjustment, with a maximum value of 4000 meaning multiply by 4. I understand that George Woltman is working on this (as well as numerous other things of course) for the new version of the software.

[tha]

Originally Posted by ckdo

http://v5www.mersenne.org/report_ll/?exp_lo=33600000&exp_hi=33605000 lists 33600571, 33601693, 33602077, 33602111 and 33602141. I'd call that "checked in".

Definitely. I had written a program that used the old files listed on the (v4) status page and compiled a report. I had no free time at hand, but I will soon write a new program based on the new report and bring the list up to date. That will allow me to post some new exponents available as well.

[Batalov]

Just in case - updated ETAs:

Originally Posted by Batalov

Test=33600643,67,1 ETA April 10
Test=33600727,67,1 ETA April 10
Test=33600877,67,1 ETA March 28
Test=33601109,67,1 ETA March 28
Test=33601201,68,1 ETA April 10

At the same time, two questions --
1. Received a couple of these, though:

Iteration: 22011501/33600877, ERROR: ROUND OFF (0.49609375) > 0.40
Continuing from last save file.

I guess, since it honestly continues from the last save file, this is just a warning (of course, if it's just one or two), yes/no?

Thank you. Yes, I had a problem with voltage at that moment. I've turned ON "round off checking" for hours after that. there were no more errors, ever.

2. http://mersenne.org/ips/manualtests.html doesn't really work. E-mail to GW, then, when done?

[S485122]

concerning ROUND OFF errors

Iteration: 22011501/33600877, ERROR: ROUND OFF (0.49609375) > 0.40
Continuing from last save file.

you should look in the helpfiles (stress.txt), this forum or the Mersenne Wiki. It points to prime95 instability, especially if you get more than one on the same machine.

Jacob

[tha]

Originally Posted by Batalov

2. http://mersenne.org/ips/manualtests.html doesn't really work. E-mail to GW, then, when done?

That is what I do.

[cheesehead]

Originally Posted by Batalov

2. http://mersenne.org/ips/manualtests.html doesn't really work.

"Manual Test Results Check In" works just fine for me. Results I reported there ten days ago now show up in the V5 reports. (There is a some-number-of-days delay because the automatic V4-to-V5 bridge is still being developed, so George has to send updates fron V4 to V5 manually, but that's just a batch job AFAIK and it does work.)

What, exactly, doesn't work for you?

[Batalov]

Originally Posted by cheesehead

What, exactly, doesn't work for you?

I've m-c-in'ed some factors and they didn't show up for a few days. Then I've emailed them - and they showed up immediately. So, I agree, it wasn't a scientific test. I should have not emailed, then iff they would show up in reports - I would have known that the manual check-in works.

I will try it again tonight. M33600877 and M33601109 will be just-in-time for this "test" done on my home comp. And this time I will deliberately not email them, only manually check-in.

And another test - I will submit my P-1 off-line tests right now (I run them on a server that doesn't email but has 32G of memory, so stage 2 is very deep) - and then I will observe the "*" disappearing in V4 reports... or not. This is a better test of man-c-in than factors or LL (which indeed have to pass the manual bridge to be seen, which introduces a delay which we cannot guesstimate no matter how hard we try)....

P.S. (c-in output below) .....nope, these transactions didn't do anything to the user report - which is I guess the fastest way to query the V4 database. I will wait for the next hours' status.txt webpage, maybe it will change?

Accepted [Line 1]: UID: Batalov/Opteron2218, M42736483 no factor from 2^68 to 2^69, Wd2: 9A441C6D
Accepted [Line 2]: UID: Batalov/Opteron2218, M42791101 no factor from 2^68 to 2^69, Wd2: 9AB81CC2
Accepted [Line 3]: UID: Batalov/Opteron2218, M36086689 no factor from 2^67 to 2^68, Wd2: 5A6AF025
Accepted [Line 4]: UID: Batalov/Opteron2218, M35975509 no factor from 2^67 to 2^68, Wd2: 595EEF70
----- rate regulated at 0.50 Hz -----
Accepted [Line 5]: UID: Batalov/Opteron2218, M42736483 completed P-1, B1=495000, B2=9281250, Wd2: 512B5890
Thank you for using GIMPS IPS manual testing forms.
Any transactions shown above are effective immediately.

[mdettweiler]

Originally Posted by Batalov

I've m-c-in'ed some factors and they didn't show up for a few days. Then I've emailed them - and they showed up immediately. So, I agree, it wasn't a scientific test. I should have not emailed, then iff they would show up in reports - I would have known that the manual check-in works.

I will try it again tonight. M33600877 and M33601109 will be just-in-time for this "test" done on my home comp. And this time I will deliberately not email them, only manually check-in.

And another test - I will submit my P-1 off-line tests right now (I run them on a server that doesn't email but has 32G of memory, so stage 2 is very deep) - and then I will observe the "*" disappearing in V4 reports... or not. This is a better test of man-c-in than factors or LL (which indeed have to pass the manual bridge to be seen, which introduces a delay which we cannot guesstimate no matter how hard we try)....

P.S. (c-in output below) .....nope, these transactions didn't do anything to the user report - which is I guess the fastest way to query the V4 database. I will wait for the next hours' status.txt webpage, maybe it will change?

Accepted [Line 1]: UID: Batalov/Opteron2218, M42736483 no factor from 2^68 to 2^69, Wd2: 9A441C6D
Accepted [Line 2]: UID: Batalov/Opteron2218, M42791101 no factor from 2^68 to 2^69, Wd2: 9AB81CC2
Accepted [Line 3]: UID: Batalov/Opteron2218, M36086689 no factor from 2^67 to 2^68, Wd2: 5A6AF025
Accepted [Line 4]: UID: Batalov/Opteron2218, M35975509 no factor from 2^67 to 2^68, Wd2: 595EEF70
----- rate regulated at 0.50 Hz -----
Accepted [Line 5]: UID: Batalov/Opteron2218, M42736483 completed P-1, B1=495000, B2=9281250, Wd2: 512B5890
Thank you for using GIMPS IPS manual testing forms.
Any transactions shown above are effective immediately.

I tried using the manual check in page a month or two ago for some TF results, but they didn't show up anywhere, even after wating about two days. So, I emailed George, and he said that the manual check-in pages do NOT work, but that instead results should just be emailed to him.

[cheesehead]

Originally Posted by Anonymous

I tried using the manual check in page a month or two ago for some TF results, but they didn't show up anywhere, even after wating about two days.

Ever since PrimeNet V5 testing started, resources like George's time have been diverted away from our usual V4 stuff. Better adjust your expectations on turnaround time for results publication.

Are you aware that new versions of files like nofactor.zip have not been posted since February 4? This is _not_ because of any failure in manual test results check-in; it's a reporting absence.

I can't speak for personal account reports, because I don't use them.

So, I emailed George, and he said that the manual check-in pages do NOT work,

There may have been short periods of time when that was true, but it's not generally so. AFAIK the last time it was down for a long time was well over a year ago, maybe two or three years ago.

I'm referring to times when mersenne.org was up, but manual check-in was down. Might George have been referring to a general server outage, rather than a manual-results-only failure?

but that instead results should just be emailed to him.

I'm trying to save George a bit of work by encouraging folks to use the manual results check-in.

[cheesehead]

Originally Posted by Batalov

And this time I will deliberately not email them, only manually check-in.

Please allow at least two weeks after your checkin before concluding that the failure is in it rather than in the reporting.

And another test - I will submit my P-1 off-line tests right now (I run them on a server that doesn't email but has 32G of memory, so stage 2 is very deep) - and then I will observe the "*" disappearing in V4 reports... or not. This is a better test of man-c-in than factors or LL (which indeed have to pass the manual bridge to be seen, which introduces a delay which we cannot guesstimate no matter how hard we try)....

My manual-checked-in P-1 results showed up in V5 reports at the same time that my manual-checked-in factors showed up on V5. So I'm not sure what you mean about a difference in P-1 vs. factors or LL.