feature request work assignments
 

Another feature request:

Added to the following list:

0 - Whatever makes most sense
2 - Trial factoring
100 - First time primality tests
101 - Double-checking
102 - World record primality tests
4 - P-1 factoring
104 - 100 million digit primality tests
1 - Trial factoring to low limits
5 - ECM on small Mersenne numbers
6 - ECM on Fermat numbers

the option:

X - Manual work assignments.

[QUOTE=tha;367757]
X - Manual work assignments.[/QUOTE]So, you want this to signal that
1) PrimeNet should not assign anything automatically, and
2) neither the client nor PrimeNet should care about "Days of work to queue up"?
Anything else?

[QUOTE=cheesehead;367758]So, you want this to signal that
1) PrimeNet should not assign anything automatically, and
2) neither the client nor PrimeNet should care about "Days of work to queue up"?
Anything else?[/QUOTE]

I would like to be able to set some cores to manual assignments, and some cores running primenet assigned tasks. Currently I do that by tweaking the worktodo file but it is tedious to prevent mprime (running in the background started at boot) from dumping the assignments and reserving undesired assignments.

And how will you give "manual assignments" to the manual cores, without playing with the worktodo? Why can't you start two copies of P95/mprime, one with "auto" assignments, on some cores, and another one with "NoMoreWork=1" (with all the other things, days to queue, etc, which will forbid p95 of "dropping" assignments), and use worktodo.add, so you don't need to play with editor?

Actually, as cheesehead discretely pointed, I can't see any reasonable scenario where you will need P95/mp to... automatically be aware of the fact that some assignments are operated... manually....

Even in "unreasonable" scenarios, a batch file will solve it (like stop prime while the batch is running, the batch waits two seconds, then locking worktodo, replacing the lines, etc).

Dumping assignments is not related to the fact that they were been assigned manually or not, but to other "queuing" settings, or to the necessity of the assignment. If P95 considers that you are wasting time, the assignment is dropped. For example, you can't TF M1277 to 65 bits, it will tell you to use ECM, regardless of what settings you try for it.

I learned myself not too long ago that you must stop Prime95 before editing worktodo.txt.

If you stop it, then add your things, then keep going, it will do whatever worktodo.txt suggests. Apparently, though, it will stop you no matter what if a certain assignment makes no sense, which, if true, should at least be optional.

The whole idea of this project is to allow people to do what they want. What if I really really want to TF M1277 to 65 bits? A checkbox somewhere would allow someone who's taken the time to look to allow "dumb work" but would stop any accidental "dumb work".

Maybe we are dumb but that suggestion for "X" for manual work makes a lot of sense to us.

While it is possible to run two copies of the client, you should not have to.

:unsure:

[QUOTE=Xyzzy;367805]While it is possible to run two copies of the client, you should not have to.

:unsure:[/QUOTE]

Yeah, that sounds legit, too.

You are spoiled by comfort :razz:

Is his problem that he doesn't want to "edit the files by hand"? Then use "worktodo.add". Anyhow, you [U]have to[/U] add "manual" assignments somehow. That is why they are "manual": you have to [U]manually[/U] reserve them from somewhere (PrimeNet, gpu72, etc), then [U]manually[/U] copy them to your computer, in "some file". Be this file called "worktodo.add", [B][U]and you are done. [/U][/B]

Is his problem that assignments are "deleted"? that is because he either does not use the program right (i.e. editing the worktodo.txt during P95 is not stopped, exited, etc), or he does not have enough room (not using the right options in the ini files, etc). Don't edit the files when the program runs, add proper keys to your lines (if you assigned them legally it is no problem!, otherwise use "N/A", don't let the field empty), use "MaxExponents=100" and "UnreserveDays=365" in prime.txt, etc.

With or without the "x" option, what you do when the "manual work" finishes? Oh, you are vigilant and watch for that not to happen, by... well... [U]manually[/U] assigning stuff from gimps, aren't you? What the difference would be with the "x" option? Still need to do [U]exactly[/U] the same stuff. You will still need to manually queue the manual work. That is why is manual.

worktodo.add solves all your problems, you can assign manual work to any core you want, anytime you want. The only thing left may be the fact that you want the "new" manual work to be done "right now", and not queued.

Well, for that you still need to stop/exit P95, and edit the file. Maybe the right "feature request" is that one to have a "[B][U]worktodo.now[/U][/B]" (or some undoc ini switch, to add the worktodo in front of the specified worker, and not to queue it. For that, you will still need to [B][U]stop[/U][/B] the program (the worker, in fact), to save what he was doing (for a later resume) and start your new, manual, work. It is 2 clicks away.

Maybe I am stupid[SUP](TM)[/SUP], but I don't see why George would waste his time for that, to implement futile things for lazy guys. I mean, I am lazy too, but he better implement other things more important related to the speed, stability, etc, of the program. Honestly.

[QUOTE=LaurV;367827]You are spoiled by comfort :razz:

[/QUOTE]

:goodposting:

We all are... :sad:

Luigi

I have two cores working on assignments of type

Pminus1=N/A,1,2,11031509,-1,250000,5000000,65

If I put them in worktodo.add mprime will empty the file and not transfer them to worktodo.txt

If I add them manually to worktodo.txt and reboot my machine the self assigned work will nicely show up using mprime -m option 3 'Status'. Also worktodo.txt will usually be unmodified. But a day later half or more (weird, not all) of these lines will be removed and replaced by a server assignment of the type specified in 'Preferences'.

Any solution to this problem I would like to see, regardless of which form. If someone can tell me how to do that with the existing software I am fine with that.

[QUOTE=tha;367856]I have two cores working on assignments of type

Pminus1=N/A,1,2,11031509,-1,250000,5000000,65

If I put them in worktodo.add mprime will empty the file and not transfer them to worktodo.txt
[/QUOTE]
Are you sure you used the right format? The Worktodo.add has to follow the same format as the .txt file, i.e. must have [Worker #1] [Worker #2] etc lines.

I just copied your line, put it 30 times in a file, split it in four (8+7+10+5) by inserting enters, added the [Worker #1], etc lines, saved as worktodo.add, this is what I have in worktodo.txt right now, after few minutes:

[CODE][Worker #1]
DoubleCheck=8C5331400<snip>46DC0A531E5B6,31439957,70,1
DoubleCheck=9C96E02250<snip>12F167D189,31439959,70,1
Pminus1=N/A,1,2,11031509,-1,250000,5000000,65
Pminus1=N/A,1,2,11031509,-1,250000,5000000,65
Pminus1=N/A,1,2,11031509,-1,250000,5000000,65
Pminus1=N/A,1,2,11031509,-1,250000,5000000,65
Pminus1=N/A,1,2,11031509,-1,250000,5000000,65
Pminus1=N/A,1,2,11031509,-1,250000,5000000,65
Pminus1=N/A,1,2,11031509,-1,250000,5000000,65
Pminus1=N/A,1,2,11031509,-1,250000,5000000,65

[Worker #2]
DoubleCheck=0749B425AA0BA<snip>97C8C1A7,31449931,70,1
DoubleCheck=DC15DC8CCDCA19478AE36A569A542CCA,31450823,70,1
Pminus1=N/A,1,2,11031509,-1,250000,5000000,65
Pminus1=N/A,1,2,11031509,-1,250000,5000000,65
Pminus1=N/A,1,2,11031509,-1,250000,5000000,65
Pminus1=N/A,1,2,11031509,-1,250000,5000000,65
Pminus1=N/A,1,2,11031509,-1,250000,5000000,65
Pminus1=N/A,1,2,11031509,-1,250000,5000000,65
Pminus1=N/A,1,2,11031509,-1,250000,5000000,65

[Worker #3]
DoubleCheck=8D7E74<snip>78B49203,31436227,70,1
DoubleCheck=1E1CE9D09<snip>D404F22D670A,31436269,70,1
Pminus1=N/A,1,2,11031509,-1,250000,5000000,65
Pminus1=N/A,1,2,11031509,-1,250000,5000000,65
Pminus1=N/A,1,2,11031509,-1,250000,5000000,65
Pminus1=N/A,1,2,11031509,-1,250000,5000000,65
Pminus1=N/A,1,2,11031509,-1,250000,5000000,65
Pminus1=N/A,1,2,11031509,-1,250000,5000000,65
Pminus1=N/A,1,2,11031509,-1,250000,5000000,65
Pminus1=N/A,1,2,11031509,-1,250000,5000000,65
Pminus1=N/A,1,2,11031509,-1,250000,5000000,65
Pminus1=N/A,1,2,11031509,-1,250000,5000000,65

[Worker #4]
DoubleCheck=2BA61<snip>3D92D647DC5E996,31438213,70,1
DoubleCheck=22B518AD<snip>E7940DDCF1,31450897,70,1
Pminus1=N/A,1,2,11031509,-1,250000,5000000,65
Pminus1=N/A,1,2,11031509,-1,250000,5000000,65
Pminus1=N/A,1,2,11031509,-1,250000,5000000,65
Pminus1=N/A,1,2,11031509,-1,250000,5000000,65
Pminus1=N/A,1,2,11031509,-1,250000,5000000,65
[/CODE]

and I am sure that, if I let it, it will do the same job 30 times, without unreserving anything. And yes, I have "MaxExponents=100" and "UnreserveDays=365" in prime.txt, etc.

If mprime behaves differently, than it is a bug.