mersenneforum.org  

Go Back   mersenneforum.org > Great Internet Mersenne Prime Search > News

Reply
 
Thread Tools
Old 2020-07-21, 09:04   #111
M344587487
 
M344587487's Avatar
 
"Composite as Heck"
Oct 2017

2·313 Posts
Default

Quote:
Originally Posted by Prime95 View Post
Both gpuowl and prime95 allow specifying a separate directory for the residues. No scripts necessary.
Nice.
Quote:
Originally Posted by preda View Post
That would work like this with gpuowl:
1. specify -tmpDir to a non-existent folder. GpuOwl will check whether the folder exist, and log a warning if it doesn't exist, but accept it nevertheless. (this is the planned behavior, not implemented)

2. because the residues can't be written to the -tmpDir location, they'll be kept in RAM until the proof is generated.

The above technique is dangerous -- in the case of a power cut, or process kill/crash, the residues are gone and the proof can't be generated.
Not requiring the user to make a tmpfs is a neat feature.
Quote:
Originally Posted by retina View Post
I agree.

Perhaps make it "-tmpdir -" or similar if you want to use RAM.

I also hope that tmpDir and tmpdir are considered the same.
Dash is a good idea, a special case id that's not likely to be confused with its normal use signifying stdin/stdout.
M344587487 is online now   Reply With Quote
Old 2020-07-21, 09:23   #112
preda
 
preda's Avatar
 
"Mihai Preda"
Apr 2015

4B016 Posts
Default

Quote:
Originally Posted by kruoli View Post
That seems somewhat counterintuitive. When passing a non-existing folder, one would expect the program to create it.
Guess what -- that's what I thought too, and that's how I implemented it initially.

But: George is pointing tmpdir to a NFS mount. When the NFS is not mounted, he does not want a folder with the same name as the NFS mount to be "created under the hood", because that messes up with the mount.

So in the end, I think the simplest behavior is: the user is responsible for creating the tmpdir. The program will warn if it's not there, but will not create it, and will not exit if not existing.

For the other folders, such as -dir or -pool, the program will require they exist otherwise won't start. This prevents this situation: used does a typo, the program creates the new folder and writes data there, the user is confused "where are those files".
preda is online now   Reply With Quote
Old 2020-07-21, 09:27   #113
retina
Undefined
 
retina's Avatar
 
"The unspeakable one"
Jun 2006
My evil lair

25·179 Posts
Default

Quote:
Originally Posted by preda View Post
This prevents this situation: used does a typo, the program creates the new folder and writes data there, the user is confused "where are those files".
I see your motivation. But I don't like it when programs try to guess my typos and "protect" me from them. User gets what user asks for. If user effs-up then user is to blame, no one else.

If your NFS mount isn't mounted, then mount it first before telling a program to use it.
retina is offline   Reply With Quote
Old 2020-07-21, 09:31   #114
preda
 
preda's Avatar
 
"Mihai Preda"
Apr 2015

24×3×52 Posts
Default

Quote:
Originally Posted by Prime95 View Post
Primenet's first official number verified by proof: 97009879

This went through all the steps. Prime95 PRP test with proof generation, file upload by separate program, server proof processor manually launched exponent marked awailable for certification, prime95 got a certification assignment, prime95 ran and uploaded the certification.

Not close to being ready for everyday use. Error handling is far from robust, the proof file uploader is broken again, server proof processor needs to be launched automatically, prime95 settings for resource management need inventing. And on and on.
One minor note on the UI: the res64 residue that is displayed in green is not actually verified by the proof. The proof only verifies the boolean status, composite / probable-prime. IMO, once an exponent is verified through proof as composite, the old res64 can be de-emphasized. Not removed, somebody may wish to do a classic double-check, but probably kept in the "history" section only, not in the green "verified" section.
preda is online now   Reply With Quote
Old 2020-07-21, 09:32   #115
preda
 
preda's Avatar
 
"Mihai Preda"
Apr 2015

120010 Posts
Default

Quote:
Originally Posted by retina View Post
I see your motivation. But I don't like it when programs try to guess my typos and "protect" me from them. User gets what user asks for. If user effs-up then user is to blame, no one else.

If your NFS mount isn't mounted, then mount it first before telling a program to use it.
Yes I see, but apparently George's NFS has a tendency towards autonomous de-mounting or something like that :)

Last fiddled with by preda on 2020-07-21 at 09:33
preda is online now   Reply With Quote
Old 2020-07-21, 09:40   #116
retina
Undefined
 
retina's Avatar
 
"The unspeakable one"
Jun 2006
My evil lair

25×179 Posts
Default

Quote:
Originally Posted by preda View Post
Yes I see, but apparently George's NFS has a tendency towards autonomous de-mounting or something like that :)
My suggestion is by default to never create folders. Always check for existence, and fail upon not-found. But, if there is a switch, say, -createfolders, then go ahead and create everything.
retina is offline   Reply With Quote
Old 2020-07-21, 09:42   #117
preda
 
preda's Avatar
 
"Mihai Preda"
Apr 2015

24·3·52 Posts
Default

Quote:
Originally Posted by retina View Post
My suggestion is by default to never create folders. Always check for existence, and fail upon not-found. But, if there is a switch, say, -createfolders, then go ahead and create everything.
OK makes sense, uniform behavior as well. I'll look in that direction. And the dash "-" to indicate "not on disk".
preda is online now   Reply With Quote
Old 2020-07-21, 17:22   #118
Prime95
P90 years forever!
 
Prime95's Avatar
 
Aug 2002
Yeehaw, FL

23×311 Posts
Default

Quote:
Originally Posted by preda View Post
Yes I see, but apparently George's NFS has a tendency towards autonomous de-mounting or something like that :)
Yes, see my "NIC gone nuts" thread. One errant machine drags the whole network down.

Sucks for me, but good for testing. If it weren't for this blinkered machine I would not have implemented "emergency memory" in prime95. If prime95 cannot write proof residues to disk, it saves them in emergency memory (1GB by default) and continues computing. This gives me a day or two to get the NFS back on line.
Prime95 is offline   Reply With Quote
Old 2020-07-22, 04:13   #119
Prime95
P90 years forever!
 
Prime95's Avatar
 
Aug 2002
Yeehaw, FL

23×311 Posts
Default

Development has been going at breakneck pace. I probably shouldn't, but here goes: A pre-pre-beta.

If you have messages waiting to be sent in prime.spl, make sure those are sent before trying these executables.

Windows is now built with VS2019 instead of MSVC 2005, so there may be unexpected issues there.

Documentation is not up-to-date. These programs cannot upload proof files. So, you'll have to collect them for later uploading.

Linux 64-bit

Windows 64-bit
Prime95 is offline   Reply With Quote
Old 2020-07-22, 04:21   #120
Prime95
P90 years forever!
 
Prime95's Avatar
 
Aug 2002
Yeehaw, FL

23×311 Posts
Default

Please don't use the above for PRP cofactor work. I haven't changed the server to handle this yet.
Prime95 is offline   Reply With Quote
Old 2020-07-24, 09:55   #121
preda
 
preda's Avatar
 
"Mihai Preda"
Apr 2015

24·3·52 Posts
Default

Quote:
Originally Posted by Prime95 View Post
These programs cannot upload proof files. So, you'll have to collect them for later uploading.
A light python script for proof upload that might work is available in GpuOwl's sources,
see here: https://mersenneforum.org/showpost.p...16&postcount=5
preda is online now   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
Your help wanted - Let's buy GIMPS a KNL development system! airsquirrels Hardware 313 2019-10-29 22:51
Is GMP-ECM still under active development? mathwiz GMP-ECM 0 2019-05-15 01:06
LLR 3.8.6 Development version Jean Penné Software 0 2011-06-16 20:05
LLR 3.8.5 Development version Jean Penné Software 6 2011-04-28 06:21
LLR 3.8.4 development version is available! Jean Penné Software 4 2010-11-14 17:32

All times are UTC. The time now is 07:34.

Wed Sep 23 07:34:09 UTC 2020 up 13 days, 4:45, 0 users, load averages: 2.83, 2.53, 2.32

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2020, Jelsoft Enterprises Ltd.

This forum has received and complied with 0 (zero) government requests for information.

Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.
A copy of the license is included in the FAQ.