CADO-NFS Data Harvesting

[charybdis]

Originally Posted by EdH

How does this compare?

I had less duplication with 12M, but I can't say how time turned out. It seems to be a bit longer than a c163, as expected. You think I should increase qmin some more and/or take mfb1 down a bit?

This is actually a bit slower than your previous c165 from post 57, which had the same params except qmin=7M and strategy 2. My comment (in a PM, I think) about raising qmin was intended for the small c17x numbers, not c160-165. Granted, strategy 2 is probably responsible for some of the difference.

For 2LP at c165, I'd drop mfb1 to 61 and bump lim1 up to something like 80M to compensate. Also ncurves1 back up to 25. Let's wait to see what VBCurtis thinks.

[VBCurtis]

Originally Posted by charybdis

This is actually a bit slower than your previous c165 from post 57, which had the same params except qmin=7M and strategy 2. My comment (in a PM, I think) about raising qmin was intended for the small c17x numbers, not c160-165. Granted, strategy 2 is probably responsible for some of the difference.

For 2LP at c165, I'd drop mfb1 to 61 and bump lim1 up to something like 80M to compensate. Also ncurves1 back up to 25. Let's wait to see what VBCurtis thinks.

I think all of this is accurate: 58/61 for mfb's sounds right, with 59/61 also worth a test. Ditto for doubling lim1 for 2LP compared to 3LP.
I'm finding best speed/duplicate compromise when Q-max is 6 to 7 times Q-min. If Q-max exceeds 8 * Q-min, I raise Q-min accordingly on the params file for that size.

For 31-bit LP and no 3LP, I'd use ncurves of 21 and 25, but up or down a couple seems to not matter.

Edit: Try lambda0 = 1.86 for mfb0 = 58, and 1.89 for mfb0 = 59. That should improve yield.

[EdH]

Here's the latest, a c171:

N = 902... <171 digits>
tasks.I = 14
tasks.lim0 = 65000000
tasks.lim1 = 40000000
tasks.lpb0 = 31
tasks.lpb1 = 31
tasks.qmin = 15000000
tasks.sieve.mfb0 = 58
tasks.sieve.mfb1 = 89
tasks.sieve.ncurves0 = 19
tasks.sieve.ncurves1 = 10
tasks.sieve.qrange = 5000
Polynomial Selection (size optimized): Total time: 979823
Polynomial Selection (root optimized): Total time: 9797.37
Lattice Sieving: Total time: 1.25403e+07s (all clients used 4 threads)
Lattice Sieving: Total number of relations: 225750239
Found 136126593 unique, 61871366 duplicate, and 0 bad relations.
cownoise Best MurphyE for polynomial is 2.98902086e-13

[VBCurtis]

Let me know what size you plan to run next; I'll do my first batch of las test-sieving on that size so we can make progress on these settings. I plan to test a bunch of settings, and then have you run a full job on the ones I think are fastest.
There remain good reasons to test full jobs- for instance, test-sieving doesn't indicate just how many relations will be needed when I change LP bounds. Our experience provides reasonable estimates, but when I'm trying to eke out 5% more speed an estimation error of 5% can swamp whatever speed I think I'm finding.

[EdH]

Here's the latest c167 (from AS 3408):

N = 336... <167 digits>
tasks.I = 14
tasks.lim0 = 60000000
tasks.lim1 = 80000000
tasks.lpb0 = 31
tasks.lpb1 = 31
tasks.qmin = 7000000
tasks.sieve.lambda0 = 1.89
tasks.sieve.mfb0 = 59
tasks.sieve.mfb1 = 61
tasks.sieve.ncurves0 = 21
tasks.sieve.ncurves1 = 25
tasks.sieve.qrange = 5000
Polynomial Selection (size optimized): Total time: 515057
Polynomial Selection (root optimized): Total time: 31081.2
Lattice Sieving: Total time: 8.02481e+06s (all clients used 4 threads)
Lattice Sieving: Total number of relations: 226666329
Found 132454195 unique, 75257815 duplicate, and 0 bad relations.
cownoise Best MurphyE for polynomial is 5.68871761e-13

[EdH]

c170:

N = 491... <170 digits>
tasks.A = 28
tasks.lim0 = 80000000
tasks.lim1 = 60000000
tasks.lpb0 = 32
tasks.lpb1 = 32
tasks.qmin = 11000000
tasks.sieve.adjust_strategy = 2
tasks.sieve.lambda0 = 1.91
tasks.sieve.lambda1 = 2.81
tasks.sieve.mfb0 = 61
tasks.sieve.mfb1 = 90
tasks.sieve.ncurves0 = 13
tasks.sieve.ncurves1 = 8
tasks.sieve.qrange = 5000
Polynomial Selection (size optimized): Total time: 1.38663e+06
Polynomial Selection (root optimized): Total time: 33979.4
Lattice Sieving: Total time: 7.0605e+06s (all clients used 4 threads)
Lattice Sieving: Total number of relations: 337852877
Found 249395183 unique, 88457694 duplicate, and 0 bad relations.*
cownoise Best MurphyE for polynomial is 3.82979404e-13

* Not all duplicates were removed prior to Msieve filtering due to overrunning the DIM value. When this finally finishes LA/SR, I'll rerun remdeps4 and get a more valid count.

[VBCurtis]

This job made a matrix of 16M size? Yuck! 2LP params yield a matrix half this size.

Either I am badly mistaken that 32LP jobs don't yield much larger matrices than 31LP, or I'm learning that 3LP params yield much larger matrices. Bright side is that either way, I am learning something!

We don't need the extra yield from 32LP here; next time you have a C170ish we'll use 31LP and drop mfb1 by 1 also- so 31/61 and 31/89. Dunno what number of relations to target.