New LLR - Page 2

2004-04-09, 14:40

Quote:

Originally Posted by Kosmaj

thommy, what cpu are you using?

i used my 2400+ athlon, and windows xp.

ET_ · 2004-04-09, 18:13

Quote:

Originally Posted by ET_

I'm having troubles with 3*2⁹⁹⁸⁹⁷⁵-1... It's the third time that the program crashes on this value

Luigi

Again. I'm forced to use old LLR on this number.

And I found out that it works at the same speed than new LLR.

Luigi

paulunderwood · 2004-04-09, 19:43

Using IBDWT : Mersenne fftlen = 57344, Proth fftlen = 114688, Used fftlen = 65536
[Fri Apr 09 20:40:11 2004]
3*2^998975-1 is not prime. Res64: 7EE98F75BA3D8F97 Time : 2529.931 sec.

This was run with the new Windoze LLR on a P4 @ 2.53GHz with no problems. Did you use linux? Was it an Athlon? If neccessary I can repeat the test on an Athlon under Mandrake 9.2.

ET_ · 2004-04-10, 00:48

Quote:

Originally Posted by paulunderwood

Using IBDWT : Mersenne fftlen = 57344, Proth fftlen = 114688, Used fftlen = 65536
[Fri Apr 09 20:40:11 2004]
3*2^998975-1 is not prime. Res64: 7EE98F75BA3D8F97 Time : 2529.931 sec.

This was run with the new Windoze LLR on a P4 @ 2.53GHz with no problems. Did you use linux? Was it an Athlon? If neccessary I can repeat the test on an Athlon under Mandrake 9.2.

I ran the number on a Celeron under Windoze 98 SE, while Prime95 was running in background. Both programs were set with priority 8.

Luigi

Kosmaj · 2004-04-12, 09:05

Jean Penn announced a revised version of LLR fixing the Res64=0 bug described above and some others. Proth mode initialization was buggy too, and he calls it "more serious" than the Res64 bug. I wonder does it mean that some primes could have been missed and that we have to re-run our tests or not?

2004-04-09, 19:43	#28
paulunderwood Sep 2002 Database er0rr 2²×941 Posts	Using IBDWT : Mersenne fftlen = 57344, Proth fftlen = 114688, Used fftlen = 65536 [Fri Apr 09 20:40:11 2004] 32^998975-1 is not prime. Res64: 7EE98F75BA3D8F97 Time : 2529.931 sec. This was run with the new Windoze LLR on a P4 @ 2.53GHz with no problems. Did you use linux? Was it an Athlon? If neccessary I can repeat the test on an Athlon under Mandrake 9.2. Last fiddled with by paulunderwood on 2004-04-09 at 19:44*

2004-04-12, 09:05	#30
Kosmaj Nov 2003 E26₁₆ Posts	Revised version of LLR announced Jean Penn announced a revised version of LLR fixing the Res64=0 bug described above and some others. Proth mode initialization was buggy too, and he calls it "more serious" than the Res64 bug. I wonder does it mean that some primes could have been missed and that we have to re-run our tests or not?