v25.7 Destroys PrimeNet v4 work-in-progress

[Freightyard]

I've been running v24 on PrimeNet v4 until last night when I upgraded to v25.7. I had several exponents at >90% completion after weeks of progress that were wiped-out by v25.7. Instead of continuing the work in progress, v25.7 got all new exponents from PrimeNet v5. Apparently, PrimeNet v5 does not recognize exponents checked out by v4 and refused to continue those; instead, it handed out new work.

HOWEVER, the work files (i.e., "q7I03833", etc.) are still in tact!

Is there any way to feed this files back into the v24 software to complete them?

"worktodo.txt" was destroyed by v25.7. What is the syntax for this file to get v24 to finish working on the exponents that are almost complete?

[starrynte]

i simply pasted the worktodo.ini stuff into worktodo.txt, then 25.7 somehow converted into its format, upon which i just moved the assignments around in worktodo.txt to its appropriate worker.

[rx7350]

**IF** you made a backup copy of your entire PRIME95 folder before you upgraded, copy/paste the 'Pxxxxxxx' files into the folder you're running V25.73 from. Make the appropriate entrie(s) in the worktodo file, stop/start PRIME95, and the exponents should pick-up right where they left off in V24.14.

Worked for me.

[jinydu]

Originally Posted by Freightyard

I've been running v24 on PrimeNet v4 until last night when I upgraded to v25.7. I had several exponents at >90% completion after weeks of progress that were wiped-out by v25.7.

Why do that? It would have been much safer to just wait a little longer until your current tests finish. Then you would have little to lose when upgrading.

[petrw1]

Originally Posted by Freightyard

"worktodo.txt" was destroyed by v25.7. What is the syntax for this file to get v24 to finish working on the exponents that are almost complete?

This should help: http://www.mersenneforum.org/showthread.php?t=6860

Or this: http://www.mersenneforum.org/showthread.php?t=10877

[Freightyard]

Help please! My worktodo.txt was deleted when I upgraded to v25 and I don't want to lose work that was nearly complete! The exponent M43703833 was 99% and I still have the work files (p7I03833 & q7I03833).

A worktodo.txt example I found shows the format (v24) is:

[Worker #1]
Test=43862069,68,1

So I can replace 43862069 with my exponent, but...

What do I use for 68,1?

I believe 1 means P-1 is completed, so that stays 1, right?

I don't know what 68 means, but looking back in my prime.log (below) I see how_factored=68.0 -- Does this mean 68 is correct for my exponent?

MUCH THANKS!

Here are excerpts from my prime.log that may help:

prime.log - Getting assignment from server

Getting assignment from server
URL: http://mersenne.org/cgi-bin/pnHttp.exe?ga&4&64&4&27317&26500&user&password&COMP-NAME&18&2&64.0
RESPONSE:
pnResult=0
requestType=2
exponent=43703833
how_factored=68.0
Got assignment 000000000000000000000000029ADE19: LL M43703833

prime.log - P-1 Completed

[Wed Aug 27 17:08:13 2008 - ver 25.6]
Sending result to server: UID: user/COMP-NAME, M43703833 completed P-1, B1=525000, B2=13125000, Wd4: 6A10D203

URL: http://mersenne.org/cgi-bin/pnHttp.exe?ar&4&88&4&18240&41&user&password&COMP-NAME&43703833&0&74.5
RESPONSE:
pnResult=0

URL: http://mersenne.org/cgi-bin/pnHttp.exe?rm&4&138&4&18655&18467&user&password&COMP-NAME&UID:+user/COMP-NAME,+M43703833+completed+P-1,+B1=525000,+B2=13125000,+Wd4:+6A10D203%0A

[Mini-Geek]

Yes, the number in the slot that had "68" refers to how far it has been factored and should remain 68.

[Freightyard]

Originally Posted by Mini-Geek

Yes, the number in the slot that had "68" refers to how far it has been factored and should remain 68.

That worked. But curious, what about non-integers? Found this assignment in my prime.log file:

exponent=43151579
how_factored=68.5

I rebuilt my worktodo.txt that was lost and it's working now, but I'm worried I could screw up the results if I didn't get 68 or 68.5 or whatever correct.

Thanks! It's running again now!

[Mini-Geek]

Originally Posted by Freightyard

That worked. But curious, what about non-integers? Found this assignment in my prime.log file:

exponent=43151579
how_factored=68.5

I rebuilt my worktodo.txt that was lost and it's working now, but I'm worried I could screw up the results if I didn't get 68 or 68.5 or whatever correct.

Thanks! It's running again now!

I've never seen something being factored half past a bit level and don't see how that could be. http://v5www.mersenne.org/report_exp...&B1=Get+status shows it being factored to 68 bits and currently assigned to you for LL testing.

[cheesehead]

Originally Posted by Mini-Geek

I've never seen something being factored half past a bit level

What that is: a kludgey way George added a long time ago to record whether P-1 had been performed or not. 68.5 means TF'd to 68 _and_ P-1 has been performed.

Some reports will strip off the .5 before displaying this field's value for the TF bit level; a few may not, and apparently something in the v4->v5 conversion software didn't get the memo.