mersenneforum.org  

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

Reply
 
Thread Tools
Old 2020-07-26, 23:29   #166
Prime95
P90 years forever!
 
Prime95's Avatar
 
Aug 2002
Yeehaw, FL

3·7·11·31 Posts
Default

Quote:
Originally Posted by Uncwilly View Post
So, I have to do at least one PRP or PRP-DC on this machine. PRP-CF does not count toward that I guess. My PRP-CF-DC's are doing the VDFs.

Just changed server, you can have CERTs too.
Prime95 is offline   Reply With Quote
Old 2020-07-28, 01:04   #167
Chuck
 
Chuck's Avatar
 
May 2011
Orange Park, FL

5·173 Posts
Default The GPU72 proxy

The first CERT I received through the proxy said there was a mismatch in the downloaded hash. George redid the CERT the next day.

I turned off the use of the proxy in the PrimeNet configuration window. The next CERT I received was a successful verification. Is there really any reason to use this proxy anymore?

My TF work is done either on Colab or through MISFIT. The only downside I see is the loss of visibility of the GIMPS DC work on the GPU72 web site.
Chuck is offline   Reply With Quote
Old 2020-07-28, 01:49   #168
chalsall
If I May
 
chalsall's Avatar
 
"Chris Halsall"
Sep 2002
Barbados

2·4,657 Posts
Default

Quote:
Originally Posted by Chuck View Post
I turned off the use of the proxy in the PrimeNet configuration window. The next CERT I received was a successful verification. Is there really any reason to use this proxy anymore?
Yeah... I've been collecting data, but haven't had time to drill down yet. The Proxy is definitely causing errors; it seems on the initial assignment, rather than on the return messaging.

Other than for P-1 work, the Proxy is little more than a way to collect stats on DC and LL work nowadays.
chalsall is offline   Reply With Quote
Old 2020-07-28, 09:24   #169
kriesel
 
kriesel's Avatar
 
"TF79LL86GIMPS96gpu17"
Mar 2017
US midwest

22×11×101 Posts
Default

Quote:
Originally Posted by Prime95 View Post
Command line arguments are user_id, proof_file_name, chunk_size, bandwidth_limit

Default chunk size is 5 (MB)
Default bandwidth_limit is unlimited. I'm using 1 (Mbps)
uploader user_id proof_file_name[ chunk size[ bandwidth_limit]
in that order apparently.

There's no -h, -?, /?
kriesel is offline   Reply With Quote
Old 2020-07-28, 15:04   #170
Xyzzy
 
Xyzzy's Avatar
 
"Mike"
Aug 2002

22·41·47 Posts
Default

How compressible are the proof files?

Maybe compressing them before uploading will help?

Interesting article: https://dev.to/biellls/compression-c...d-deflate-15g1
Xyzzy is offline   Reply With Quote
Old 2020-07-28, 15:20   #171
kruoli
 
kruoli's Avatar
 
"Oliver"
Sep 2017
Porta Westfalica, DE

1001100112 Posts
Default

That got discussed around page #3 in this thread already. Unfortunately, compression would not help.
kruoli is offline   Reply With Quote
Old 2020-07-28, 16:53   #172
Uncwilly
6809 > 6502
 
Uncwilly's Avatar
 
"""""""""""""""""""
Aug 2003
101×103 Posts

100001101011012 Posts
Default

Corporate security settings are too strict for such a new version of Prime95 and the libgmp-10.dll.

Oh well.

Home machine has received, run, and turned in PRP-CF certs and has generated some for others.
Uncwilly is offline   Reply With Quote
Old 2020-07-28, 21:19   #173
ATH
Einyen
 
ATH's Avatar
 
Dec 2003
Denmark

B7C16 Posts
Default

I tried v30.1b1 on PRP CF and it generated a proof with no issues: 3344339

But when there is already a type 5 PRP CF test on the exponent and 1 extra factor is found, I think 1 test is enough to verify the cofactor right? The residue matched the previous tests, so no proof validation is needed.

Last fiddled with by ATH on 2020-07-28 at 21:20
ATH is offline   Reply With Quote
Old 2020-07-29, 17:00   #174
chalsall
If I May
 
chalsall's Avatar
 
"Chris Halsall"
Sep 2002
Barbados

221428 Posts
Default

Hey George. More data...

So, my Winblows machine has done two CERT runs in the last day. 97265881 matched, while 97282747 didn't (your run confirmed the original PRP run was good).

Neither of these went through the Proxy. Direct comms with Primenet.

Is it possible this new code path is exercising different part(s) of my kit, and I might need to service it? I /think/its history with DC'ing has been flawless.
chalsall is offline   Reply With Quote
Old 2020-07-29, 17:18   #175
kriesel
 
kriesel's Avatar
 
"TF79LL86GIMPS96gpu17"
Mar 2017
US midwest

22×11×101 Posts
Default Current status of PRP proof / cert verification rollout

It's progressing very well.
Producing prp proofs: prime95/mprime v30.1, gpuowl v6.11-312? or later
As far as I know these are the only applications which have implemented and demonstrated support.

PrimeNet API is sending Cert work to and retrieving results from suitable prime95/mpprime instances
Gpuowl's primenet.py can be used to upload proof files on systems with python installed.
There's also a manual uploader program provided by Prime95 which can be used with gpuowl proof files.
(That does not require any compiler or interpreter present on the Windows system.)

The PrimeNet server handles quite a few aspects of proof and verification. The essential function of getting verifications done are there.
Areas at mersenne.org that may benefit from updating eventually:
https://www.mersenne.org/report_recent_cleared/ does not list certs
https://www.mersenne.org/assignments/ does not have an exclude certs choice
https://www.mersenne.org/primenet/ has no info on certs, no distinction between LL & PRP
https://www.mersenne.org/account/?details=1 has no info on certs performed
https://www.mersenne.org/report_top_500/ has no cert info
https://www.mersenne.org/report_top_500_custom/?type=0 has no stats type for certs
https://www.mersenne.org/report_top_teams/ has no cert info
https://www.mersenne.org/report_top_...om/?type=10000 has no stats type for certs
There's no detailed report for certs analogous to https://www.mersenne.org/report_prp/
Cert history entries don't indicate a result https://www.mersenne.org/report_expo...7513771&full=1 for example, Architects Cubed's Cert did not match, Woltman's did match, but you can't tell from the page.
There is currently no manual CERT assignment, and that may be by design. https://www.mersenne.org/manual_assignment/

It's unclear how mersenne.ca will handle proof-based verifications.
https://www.mersenne.ca/exponent/97018073 does not indicate any cert, or prp residue near top of page, although mersenne.org shows it with a cert and prp residue from 7/22 and 7/23 respectively.

Last fiddled with by kriesel on 2020-07-29 at 17:20
kriesel is offline   Reply With Quote
Old 2020-07-29, 17:57   #176
firejuggler
 
firejuggler's Avatar
 
Apr 2010
Over the rainbow

2,417 Posts
Default

I guess that this is a 'proof' and not a cert....
well it take me about a week.
Attached Thumbnails
Click image for larger version

Name:	Screenshot_2.jpg
Views:	40
Size:	32.6 KB
ID:	22910  

Last fiddled with by firejuggler on 2020-07-29 at 17:58
firejuggler 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 09:28.

Mon Sep 28 09:28:21 UTC 2020 up 18 days, 6:39, 0 users, load averages: 1.54, 1.44, 1.38

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.