When do we stop

[azhad]

Hi,

I have worked on GIMPS for several years now. My current PC churns out good results in LL/PRP. Main processor is the work horse, an i7-3770K, water-cooled, running overclocked at 4.2GHz.

Graphics is 2 x AMD 7970 series cards - which crash my comp in 15 minutes, hence no use for GIMPS.

It appears that 100M mersenne's are so far away. My first question is when do we stop? It appears that GIMPS will never stop or reach fulfillment.

My second question is, what has been done to ensure all this work is saved forever?

[retina]

Originally Posted by azhad

My first question is when do we stop?

We will be forced to stop when we run out of exponents to test.

[kriesel]

Originally Posted by azhad

Hi,

I have worked on GIMPS for several years now. My current PC churns out good results in LL/PRP. Main processor is the work horse, an i7-3770K, water-cooled, running overclocked at 4.2GHz.

Graphics is 2 x AMD 7970 series cards - which crash my comp in 15 minutes, hence no use for GIMPS.

It appears that 100M mersenne's are so far away. My first question is when do we stop? It appears that GIMPS will never stop or reach fulfillment.

My second question is, what has been done to ensure all this work is saved forever?

p=~100M are being run now: https://www.mersenne.org/assignments...&exp1=1&extf=1
100M-DIGIT are being run now in even greater numbers: https://www.mersenne.org/assignments...&exp1=1&extf=1
Lately our rate of progress through minor milestones has shown GIMPS to advance its wavefront of exponents under first-primality-test by about 6M/year of exponent value. https://www.mersenneforum.org/showpo...73&postcount=2 Extrapolating that forward under the assumption that the number and speed of hardware engaged will compensate for the increased effort required for higher exponents growing as p log p log log p or approx p^2.1 gives an estimate of reaching p~10⁹ of about 150 years. https://www.mersenneforum.org/showpo...5&postcount=11 Under these same assumptions the 100Mdigit area appears to be about 40 years away.

Things that might disrupt the effort (grid failure, disruption of technological civilization, war, disease, etc.) could change the schedule substantially. Quantum computing might speed it up. An unexpected number theory advance might speed it up or make it moot. If a proof was achieved that there were no Mersenne primes above a certain plim, primality testing and verifying all prime exponents below plim would conclude the search.

On an individual basis, we'll stop when we run out of interest, finances and hardware to continue, mental faculty to keep the tests going and reporting in, or days of life span.

Never and forever are very long times. We're open to suggestions how to preserve data that long. Here is a list of some possible challenges to consider. https://en.wikipedia.org/wiki/Timeli...the_far_future