Oddities in large P-1 candidates

clowns789 · 2018-04-16, 05:11

Out of curiosity and audacity, I decided to look at the largest candidates within Prime95's spectrum (<595.8M). The top 3 have factors, but the two below it have had P-1 work done, albeit seemingly not reported properly. M595799933 is reported as still having P-1 in progress although it says it is completed below, and M595799947 shows as repeating some TF from 2012 last month and then running P-1 with 1M as the B1 bound, and reported twice. Is there any explanation for this in either case?

tha · 2018-04-17, 06:56

That range is so far from the server assigned ranges that all work done there is incidental.

kriesel · 2019-06-08, 17:20

There are oddities in how the P-1 programs act, particularly on gpus, apparently as a function of gpu ram, exponent, bounds, and other undetermined variables. Some runs are being made at exponents well above the assignment wavefronts. See https://www.mersenneforum.org/showpo...82&postcount=7 and https://www.mersenneforum.org/showthread.php?t=23389

2018-04-16, 05:11	#1
clowns789 Jun 2003 The Computer 401 Posts	Oddities in large P-1 candidates Out of curiosity and audacity, I decided to look at the largest candidates within Prime95's spectrum (<595.8M). The top 3 have factors, but the two below it have had P-1 work done, albeit seemingly not reported properly. M595799933 is reported as still having P-1 in progress although it says it is completed below, and M595799947 shows as repeating some TF from 2012 last month and then running P-1 with 1M as the B1 bound, and reported twice. Is there any explanation for this in either case?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
No available candidates on server	japelprime	Prime Sierpinski Project	2	2011-12-28 07:38
TF Credit oddities	Christenson	PrimeNet	2	2011-03-29 02:11
new candidates for M...46 and M48	cochet	Miscellaneous Math	4	2008-10-24 14:33
MSieve oddities....	schickel	Msieve	4	2007-04-15 21:42
Oddities of democracy	kwstone	Soap Box	3	2004-02-01 19:03

2018-04-17, 06:56	#2
tha Dec 2002 3²×97 Posts	That range is so far from the server assigned ranges that all work done there is incidental.

2019-06-08, 17:20	#3
kriesel "TF79LL86GIMPS96gpu17" Mar 2017 US midwest 2²·3·17·37 Posts	There are oddities in how the P-1 programs act, particularly on gpus, apparently as a function of gpu ram, exponent, bounds, and other undetermined variables. Some runs are being made at exponents well above the assignment wavefronts. See https://www.mersenneforum.org/showpo...82&postcount=7 and https://www.mersenneforum.org/showthread.php?t=23389