Thread: Announcements
View Single Post
Old 2021-01-24, 19:20   #1
kriesel
 
kriesel's Avatar
 
"TF79LL86GIMPS96gpu17"
Mar 2017
US midwest

3×52×79 Posts
Default Announcements

(Most content in this post will come and go and be of temporary interest at best.)


Mersenne.ca now also accepts P-1 results for exponents > 1G, from gpuowl (factors found tested lightly with v6.11-380 at 1.13-1.71G; no factor found not yet tested.) Put the result(s) in a text file and upload at https://www.mersenne.ca/bulk-factors.php

James Heinrich has created a converter for Mfactor and Factor5 multi-line results to standardized json format.
https://www.mersenne.ca/mfactor2json.php


Mprime / prime95 v30.7b9 is available for test, with some modifications to P-1, addition of Alder Lake support
Mlucas V20.1.1 (Nov 6 version) has been released, with faster P-1 stage 2, various fixes, separate help.txt file.
(For latest versions etc of leading software, see http://www.mersenneforum.org/showpos...91&postcount=2)


PRP/GEC residual error rate per test over the period 2019-08-12 to 2021-08-12, for exponents above 50M, is empirically estimated as only ~24 ppm. That's just 3 errors in over 123,000 completed tests. At least one of those errors occurred before the final-residue handling outside the GEC code was hardened against error in prime95. It's possible all 3 were from runs on code versions from before that hardening change. Even if not, that's vastly better than typical LL error rate of ~2%/primality test at the ~100M wavefront, ~20% at 100Mdigit, estimated ~90% at 1G.


Important!
If performing first primality tests, use PRP with GEC and proof generation. NO LL for first tests! Or for DC either. As of 2021-04-08, the server no longer issues LL first time test assignments.

Please switch for first tests, to PRP testing, with the excellent Gerbicz Error Check, and VDF based verification of correct completion ("proof"), wherever practical, instead of doing LL first tests. Cease beginning LL first primality tests. All new first primality tests should be PRP type, with proof generation whenever possible. PRP/GEC/proof/Cert is much more reliable, and effectively twice as fast, since it essentially eliminates the typical 2% LL error rate leading to occasional triple checks or further, and reduces verification to typically under 1% of what an LL DC requires (>99% reduction!). If the choices are between Gpuowl PRP or CUDALucas LL, the savings are even more dramatic, nearly 4:1. PRP with proof generation also provides much faster indication of system reliability (typically hours), than does LL. (Typically years, 8.2 in that example, for an LL DC to occur; examples over 9 years also exist, including: https://www.mersenne.org/report_expo...6440291&full=1
https://www.mersenne.org/report_expo...7890291&full=1
https://www.mersenne.org/report_expo...5182317&full=1
https://www.mersenne.org/report_expo...4926731&full=1
https://www.mersenne.org/report_expo...7082673&full=1
https://www.mersenne.org/report_expo...5916561&full=1
https://www.mersenne.org/report_expo...7224663&full=1
https://www.mersenne.org/report_expo...7357359&full=1
https://www.mersenne.org/report_expo...7458873&full=1
https://www.mersenne.org/report_expo...7939197&full=1
etc.
Ten years:
https://www.mersenne.org/report_expo...2197123&full=1)


Since the reliability advantage of PRP/GEC over LL also applies to double checking, most LL first tests will never have LLDC run. They will instead be verified composite more reliably and more quickly by a first time PRP/GEC/proof and cert. So in the usual >999,999 ppm case of a composite Mersenne number, an LL first primality test is a complete waste of time and computing resources. (An LL DC on an LL first test is a lesser waste of time, and much less reliable than PRP/GEC.)

Mprime/prime95 (~v30.3 or higher), and GpuOwl (~v6.11-316 and later) are available in proof-capable forms. Mlucas proof capability is planned for V20.x. (Lucas-Lehmer tests will still be used if/when PRP returns a probable prime indication for a suspected new Mersenne prime discovery, to confirm by multiple independent tests by separate hardware, software, and participants. The person reporting a PRP probable prime indication from a first test would still be regarded as the discoverer.)
In occasional sampling ~2021-01-24 of over 130 primality testing assignments in 101-102M, 38% were LL, 62% PRP. That LL frequency is much too high.
As of 2021-01-27 1540 UTC in 101-102M primality testing of 273 active assignments, 57 LL, 21%, better. For 101M-103M as of 2021-03-28 ~2010 UTC, 23 LL of 132, 17%. On 2021-04-21 ~1500 UTC, for 102M-104M, first 1000 contained only 41 LL, 4.1% On 2021-04-30 ~1800 UTC, for 102M-104M, first 1000 contained only 11 LL, 1.1%. Let's keep moving away from inefficient LL first tests, that are very slowly double-checked, toward complete conversion to the far more reliable and efficient PRP/GEC/proof/Cert that also provides very rapid verification. A check of "recently cleared" 2021-07-19 showed a mix of primality tests for exponents >100M, 12 LL (mostly elapsed time over a year) vs. 924 PRP, ~1.2%. Drive LL percentage toward 0%, PRP with proof generation toward 100%. As of 2021-08-28, https://www.mersenne.org/assignments...chk=1&excert=1 shows only PRP in 1000 queued tests. It's unclear how many will produce a proof file.
GPUs that can run either CUDALucas or Gpuowl should run PRP/GEC/proof on Gpuowl as primality tests, not LL, and not CUDALucas. (That would include all Google Colab GPU models. The Google Drive free space is sufficient.)


Punctuality
Please run assignments promptly or visit https://www.mersenne.org/workload/ to turn them back in so someone else can run them.


(This is not a discussion thread. Please do not post in this thread. Use PM or a discussion thread. Most content in this thread will be temporary.)
Attached Files
File Type: pdf prp preferable.pdf (21.8 KB, 111 views)

Last fiddled with by kriesel on 2021-11-22 at 15:57 Reason: Mfactor/factor5 converter available, mersenne.ca accepting gpuowl P-1 factors
kriesel is offline