At 08:14 07-10-98 -0700, you wrote:
>Remember, It will take much longer to test exponents in the 8 million
>range. Did you take that into consideration?
>
>I bet by the time we reach 8 million, it will take nearly 6 months to
>complete an exponent.
Computer are going faster too. More seriously, I estimate the total CPU time
in P90-years based on a linear progession of our computing capability (see
the Primenet stat page). Then I use the table on the GIMPS status page to
find out where we will be at that time.
We shall remember than the limits are approximative and ill define. By
example,
there is some people who testing number in the 6 200 000 range now and
result are coming out of the 5 900 000 range, but there is still people
testing numbers in
the 3 100 000 range. At this, the job is probably complete up to 90 percent
below
5 200 000.
Maybe a multi-level objective would be better. Say for 2001:
-Exploration factoring up to 15 000 000
-one L-L up to 10 000 000
-two L-L up to 6 000 000
-ECM complete factoring up to 100 000
Yvan Dutil
