This took 460 GPU-hours (so about ten days on two GPUs).

First I ran a calibration job to see what stage_1 value was required to get about 20 stage-1 hits per second; the answer turns out to be about 1e27. Then I ran some short ranges (1-2M and 2M-3M), took a 1% sample of the stage-1 hits and ran -nps with a high stage2_norm, sorted by output score and reran on the whole set of stage-1 hits with the threshold equal to the tenth-best output score in the sample - that is, I was aiming for about a thousand stage-2 hits. Once I knew how long the short ranges had taken to run, I ran some longer ranges (3M-17M on one GPU, 17M-30M on the other) and repeated the process.

After running -npr on the thousand-or-so stage-2 hits, I sort by E-value and note the hundredth-best as a plausible threshold for the next job.

So for 114!+1, difficulty 186.4, the stage-2 threshold I found good for the long ranges is 1.2e+25 (this actually got about 2000 hits), and the hundredth-best-E-value on both ranges was about 2.6e-14.

I've just used another 465 GPU-hours for the C179 from 139^137+137^139

This time I used stage 1 norm 2.5e26, searching from 1 to 2.6M to find 99.3M unique stage-1 hits. I was aiming to run 400 GPU-hours for 1-2M, so used the rough rule that doubling stage1-norm takes 5 times as long to get 8 times the yield to pick a stage-1 norm that took about two hours to run 1M..1M+10k; but 1M-2M ran faster than 1-1M so I extended the range that side).

Rest was routine; 163 CPU-hours to run -nps with a (relatively low) stage-2 norm of 5e23, 19.5 hours to run -npr on the 1541 survivors. Best polynomial had a score of 1.012e-13 (and a stage-2 norm of 2.39e23); hundredth-best was 8.87e-14.