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

737410 Posts
Default Announcements

(Most content in this post will come and go and be of temporary interest at best. Generally content here is newest first)

2023--2-01- Pre-Beta version of prime95

2023-01-22 There may be a bug in mprime or prime95 v30.8b1x preventing getting cert assignments under some conditions. There are definitely also some settings which turn off requesting cert work. See

2023-01-16 Google Colab VM configuration changes ~2023-01-12 break some if not all previously working GPU-use notebook sections. See

There's a web page based manual proof files uploader at

Mihai Preda has created a version of gpuowl capable of outputting P-1 savefiles compatible with mprime. See and several following posts.
(P-1 stage 1 at low incremental overhead during an earlier v7.2-x PRP run, and standalone two-stage P-1 via V6.11-3xx, are still available as options. Install different versions in different folders is a useful approach.)

There's a paper describing a generalization of GEC and VDF, which appears to allow adding the same excellent error detection and verification now available for PRP, to P-1 stage 1 computations. Discussion starts here in

Gpuowl Windows builds allowing greater range of proof power for PRP proof generation have been created & posted for V6.11-382 and V7.2-93. See the last paragraph of for more info and links to the relevant posts.

Mlucas V20.1.1 (2022-07-06 patched version) has been released.
(For latest versions etc of leading software, see

A 7-undergraduate-student team under the guidance of danc2 and tdulcet has made a start, on code to perform probable prime tests of Mersennes running on a Google Colaboratory TPU. Read more in the last paragraphs of this post. now supports reserving P-1 for OBD exponents; see now accepts P-1 factoring results from mprime/prime95 including from V30.8b14, in json format, bulk factor file upload, for exponents above 1G up to the AVX512 limit 1169M of prime95. See also paragraph beginning "Note" here. now also accepts P-1 results for 1G < exponent < 232, from gpuowl or Mlucas. (factors found tested lightly with gpuowl v6.11-380 at 1.13-1.71G; no factor found also tested.) Put the result(s) in a text file and upload at Mlucas output needs a user JSON field added before upload. For example, the bold underlined red content beginning with a comma in the following was manually added:
{"status":"F", "exponent":2147483647, "worktype":"PM1", "fft-length":125829120, "B1":1000000, "B2":100000740, "partial-stage-2":true, "factors":["242557615644693265201"], "program":{"name":"Mlucas", "version":"20.1.1"}, "timestamp":"2023-01-20 23:51:22 UTC", "user":"kriesel", "computer":"roa"}
James Heinrich has created a converter for Mfactor and Factor5 multi-line TF results to standardized json format.

If performing first primality tests, use PRP with GEC and proof generation. NO LL for first tests! Or for DC or TC of LL past first tests either. As of 2021-04-08, the server no longer issues LL first time test assignments. PRP with GEC and proof generation is more reliable and more efficient than using LL, in first test, DC, TC, etc.
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. See also this.

Please run assignments to completion promptly or visit to turn them back in so someone else can run them. Upload proof files promptly after their generation and reporting the result line.

(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, 182 views)

Last fiddled with by kriesel on 2023-02-01 at 07:30 Reason: added notes about cert work; grouped P-1 notes
kriesel is offline