mersenneforum.org PRP vs. LL (was: Changes to Recent Cleared and Recent Results)
 Register FAQ Search Today's Posts Mark Forums Read

 2019-03-14, 17:59 #1 petrw1 1976 Toyota Corona years forever!     "Wayne" Nov 2006 Saskatchewan, Canada 23×569 Posts PRP vs. LL (was: Changes to Recent Cleared and Recent Results) Was this announced? (I admit I don't follow every post) I'm not complaining; only asking. Recent Cleared: No longer reports LL/DC https://www.mersenne.org/report_recent_cleared/ Recent Results: LL/DC are now indicated with C-LL instead of C. https://www.mersenne.org/report_recent_results/ Any other changes I missed? Are more changes coming?
2019-03-14, 20:53   #2
Prime95
P90 years forever!

Aug 2002
Yeehaw, FL

1CB616 Posts

Quote:
 Originally Posted by petrw1 Was this announced? (I admit I don't follow every post) I'm not complaining; only asking. Recent Cleared: No longer reports LL/DC https://www.mersenne.org/report_recent_cleared/ Recent Results: LL/DC are now indicated with C-LL instead of C. https://www.mersenne.org/report_recent_results/ Any other changes I missed? Are more changes coming?

I did this --- for consistency.
Composite by LL was "C"
Composite by PRP was "PRP-C"
Factors found was "F", "F-PM1", "F-ECM"

They are now "result"-"method" (except F and NF where no method means TF).

Why? I'm working on changing the "what makes most sense" to PRP for v29.6 and later. To do this the web pages are being examined to make sure they are equally informative for both LL and PRP users.

What else has changed thusfar? PRP results are formatted a bit cleaner on recent results/cleared pages. There is a new stats type in the custom top 500 page. First primality tests combines LL and PRP stats. Double-checking combines LL and PRP double-check stats. I want to change the Top 500 menu choices to default to these combined stats (not done yet).

The user assignments page now tells you original and current category for PRP assignments. The ability to get Cat 0/1/2 assignments now factors in recent PRP results as well as LL results.

The main web page now reports yesterday's primality tests -- not yesterday's LL tests.

Recommendations for what else needs to be done are welcome. The goal to change the user's perspective to "I'm primality testing" away from "I'm LL testing". Casual users don't really care if it is an LL test or a PRP test that finds the next Mersenne prime.

Last fiddled with by Prime95 on 2019-03-14 at 20:55

 2019-03-14, 21:41 #3 lycorn     Sep 2002 Oeiras, Portugal 26508 Posts And what about the PRP on Mersenne cofactors type of work? I see many of them reported in the Recent Cleared Report, but I can´t figure out how to get them via the Manual Assignments.
2019-03-15, 01:30   #4
Prime95
P90 years forever!

Aug 2002
Yeehaw, FL

2·3·52·72 Posts

Quote:
 Originally Posted by lycorn And what about the PRP on Mersenne cofactors type of work? I see many of them reported in the Recent Cleared Report, but I can´t figure out how to get them via the Manual Assignments.
Try now.

 2019-03-15, 06:47 #5 lycorn     Sep 2002 Oeiras, Portugal 101101010002 Posts There they are! Thank you.
2019-03-15, 07:47   #6
srow7

Jul 2014

2·52 Posts

Quote:
 Recent Results: LL/DC are now indicated with C-LL instead of C
George, I think you optimized too much.
where are the DC results in the 50M range ?

I assume they should be displayed 5x,xxx,xxx C-LL

 2019-03-15, 10:22 #7 henryzz Just call me Henry     "David" Sep 2007 Cambridge (GMT/BST) 2·2,909 Posts The exponent status distribution doesn't mention prp tests https://www.mersenne.org/primenet/ A totals row would also be interesting(maybe also at 100M)
2019-03-15, 19:22   #8
kriesel

"TF79LL86GIMPS96gpu17"
Mar 2017
US midwest

2×3×19×43 Posts

Quote:
 Originally Posted by Prime95 ... I'm working on changing the "what makes most sense" to PRP for v29.6 and later. To do this the web pages are being examined to make sure they are equally informative for both LL and PRP users. What else has changed thusfar? PRP results are formatted a bit cleaner on recent results/cleared pages. There is a new stats type in the custom top 500 page. First primality tests combines LL and PRP stats. Double-checking combines LL and PRP double-check stats. I want to change the Top 500 menu choices to default to these combined stats (not done yet). The user assignments page now tells you original and current category for PRP assignments. The ability to get Cat 0/1/2 assignments now factors in recent PRP results as well as LL results. The main web page now reports yesterday's primality tests -- not yesterday's LL tests. Recommendations for what else needs to be done are welcome. The goal to change the user's perspective to "I'm primality testing" away from "I'm LL testing". Casual users don't really care if it is an LL test or a PRP test that finds the next Mersenne prime.
Makes sense, and I'm looking forward to the better integration of PRP in the various pages.
I note on my https://www.mersenne.org/workload/ a manual prp assignment entry now shows a negative value for Expires (days):
Code:
Kriesel has 327 assignments
Manual testing    1    83411369     PRP    0    1        2019-02-06    2019-02-06            -7
PRP assignments until now have had no expiration associated. Is this a byproduct of your recent changes?

Last fiddled with by kriesel on 2019-03-15 at 19:23

2019-03-15, 19:42   #9
Prime95
P90 years forever!

Aug 2002
Yeehaw, FL

2×3×52×72 Posts

Quote:
 Originally Posted by henryzz The exponent status distribution doesn't mention prp tests https://www.mersenne.org/primenet/
There is (kind of). The footer indicates that LL means a count of LL and PRP tests.

Quote:
 A totals row would also be interesting(maybe also at 100M)
Agreed, but that is not related to the make-PRP-testing-as-easy-as-LL upgrades.

2019-03-15, 20:06   #10
Prime95
P90 years forever!

Aug 2002
Yeehaw, FL

2×3×52×72 Posts

Quote:
 Originally Posted by srow7 George, I think you optimized too much. where are the DC results in the 50M range ? I assume they should be displayed 5x,xxx,xxx C-LL

Oops. The view that creates the recent cleared had a where clause based on the "C" and "PRP" rather than the numeric result code. Fixed.

2019-03-15, 20:09   #11
Prime95
P90 years forever!

Aug 2002
Yeehaw, FL

2·3·52·72 Posts

Quote:
 Originally Posted by kriesel PRP assignments until now have had no expiration associated. Is this a byproduct of your recent changes?
I did not change any of this. Aaron did the expiration calculations. I've been doing PRP for a while and my workload page does not have any negative values.

 Similar Threads Thread Thread Starter Forum Replies Last Post petrw1 PrimeNet 5 2018-07-09 21:14 ramgeis PrimeNet 8 2016-05-14 23:51 Primeinator PrimeNet 4 2014-10-26 00:56 garo PrimeNet 7 2014-02-03 02:53 kladner Software 20 2011-09-02 02:59

All times are UTC. The time now is 11:14.

Fri Feb 26 11:14:55 UTC 2021 up 85 days, 7:26, 0 users, load averages: 1.44, 1.26, 1.28