COMPLETE!!!! Thinking out loud about getting under 20M unfactored exponents

[petrw1]

Originally Posted by LaurV

We are, for a little while, back to tickling 11M exponents with P-1. Santa was nice this year.
Starting 11.2 and 11.0, in parallel.

Thanks

[alpertron]

Originally Posted by petrw1

Wow thanks ... I assure you the average will be much better than 51 per factor
Do I understand you are working on all of 10.0 to 10.9?

Yes, the idea is to process first the Mersenne numbers with exponents from 10,000,000 to 11,000,000 which currently have lower bounds for P-1. It appears that most of them are in the range 10.7M to 11M.

[De Wandelaar]

My new assignments are in the 22.8 range (TF 74->75)
I suppose it's ok.

[lycorn]

12.0 is getting close to finish. If it´s ok with everybody I would like to take 12.1 next.
@chalsall: if there is no objection, could you "release" it, please? Thx.

[petrw1]

Originally Posted by De Wandelaar

My new assignments are in the 22.8 range (TF 74->75)
I suppose it's ok.

I assume you meant 21.8?
Yes that is the next range of interest for TF.

[chalsall]

Originally Posted by lycorn

12.0 is getting close to finish. If it´s ok with everybody I would like to take 12.1 next. @chalsall: if there is no objection, could you "release" it, please? Thx.

It's yours. Thanks for the cycles!

[petrw1]

Should be a Christmas Present.

[Luminescence]

26.5M is finally finished and 12.3M is almost 25% done. Damn this new version is fast, 26.5M was just a bit chaotic since 30.8 threw a wrench into my plan. 12.3M seems like a sweetspot in terms of stage 1 speed for my CPUs, I think I'm gonna clean up the easier ones in 12M afterwards.

@petrw1 is your mprime/Prime95 reporting round off error with AVX-512? In my case it's the only machine that doesn't report it and uses AVX-512. It just shows

Round off: 0,

for me while FMA3 is fine.

[petrw1]

Originally Posted by Luminescence

26.5M is finally finished and 12.3M is almost 25% done. Damn this new version is fast, 26.5M was just a bit chaotic since 30.8 threw a wrench into my plan. 12.3M seems like a sweetspot in terms of stage 1 speed for my CPUs, I think I'm gonna clean up the easier ones in 12M afterwards.

@petrw1 is your mprime/Prime95 reporting round off error with AVX-512? In my case it's the only machine that doesn't report it and uses AVX-512. It just shows

Round off: 0,

for me while FMA3 is fine.

I get a round-off....her is an entire run with a few snips

[Dec 13 17:07] P-1 on M26453437 with B1=600000, B2=TBD
[Dec 13 17:07] Using AVX-512 FFT length 1440K, Pass1=192, Pass2=7680, clm=2, 8 threads
[Dec 13 17:07] Setting affinity to run helper thread 2 on CPU core #3 ... and 6 more
[Dec 13 17:16] M26453437 stage 1 complete. 1731726 transforms. Total time: 500.228 sec.
[Dec 13 17:16] Round off: 0.07780627092
[Dec 13 17:16] Conversion of stage 1 result complete. 5 transforms, 1 modular inverse. Time: 8.910 sec.
[Dec 13 17:16] With trial factoring done to 2^75, optimal B2 is 286*B1 = 171600000.
[Dec 13 17:16] If no prior P-1, chance of a new factor is 4.96%
[Dec 13 17:16] Switching to AVX-512 FFT length 1600K, Pass1=128, Pass2=12800, clm=2, 8 threads
[Dec 13 17:16] Setting affinity to run helper thread 2 on CPU core #3 ... and 6 more
[Dec 13 17:16] With trial factoring done to 2^75, optimal B2 is 256*B1 = 153600000.
[Dec 13 17:16] If no prior P-1, chance of a new factor is 4.88%
[Dec 13 17:16] Using 24831MB of memory.  D: 3570, 384x1568 polynomial multiplication.
[Dec 13 17:16] Setting affinity to run polymult helper thread on CPU core #2 ... and 6 more
[Dec 13 17:16] Round off: 0.006488925122, poly_size: 2, EB: 1.27242, SM: 2.39624
[Dec 13 17:16] Round off: 0.01016381498, poly_size: 4
[Dec 13 17:16] Round off: 0.01501164764, poly_size: 8
[Dec 13 17:16] Round off: 0.02329044813, poly_size: 16
[Dec 13 17:16] Round off: 0.03703734813, poly_size: 32
[Dec 13 17:16] Round off: 0.04735467656, poly_size: 64
[Dec 13 17:16] Round off: 0.06808246584, poly_size: 128
[Dec 13 17:16] Round off: 0.09170954395, poly_size: 256
[Dec 13 17:16] Round off: 0.08529820227, poly_size: 512
[Dec 13 17:16] Stage 2 init complete. 10136 transforms. Time: 30.614 sec.
[Dec 13 17:16] Round off: 0.1567131011
[Dec 13 17:24] M26453437 stage 2 complete. 464611 transforms. Total time: 466.070 sec.
[Dec 13 17:24] Round off: 0.1791973456
[Dec 13 17:24] Stage 2 GCD complete. Time: 5.680 sec.
[Dec 13 17:24] M26453437 completed P-1, B1=600000, B2=154081200, Wi4: #######

[petrw1]

To the table in post #1
Let me know if I missed any....my bad; I missed several...sorry.
How about now?
Thanks

[LaurV]

We just beat another milestone:

A thousand posts in the thread


And me being 2^11-1, the smallest composite mersenne** number, hihi


-----
**restricted definition