Timings reported by las can be negative
Imported issue: Initially reported by Cyril Bouvier in https://gforge.inria.fr/tracker/?group_id=2065&aid=21688
Some timings reported at the end of las are negative.
Example (with 1 special q for rsa180):
las -poly parameters/polynomials/rsa180.poly -sqside 1 -lim0 167874892 -lim1 91296860 -lpb0 31 -lpb1 32 -mfb0 62 -mfb1 99 -fb1 rsa180.roots -I 15 -q0 91296860 -nq 1 -t 1
Output excerpt:
# Time for this special-q: 13.5573s [norm 0.0120+0.0527, sieving 10.9402 (6.1498 + 0.6344 + 4.1560), factor 2.5524 (0.5248 + 2.0276)]
...
# Total cpu time 20.00s [norm 0.01+0.1, sieving 17.4 (-1.4 + 0.6 + 18.1), factor 2.6 (0.5 + 2.0)]
The error is in the first timings of the sieving part.
Edit: timings obtained with commit 93c8eea6 (current master branch of July 3rd)