Credit for trial factoring

[kladner]

I received credit of ~20 for my most recent factor in the 66M range.

[Gordon]

Originally Posted by TheMawn

[snip]

On the other hand, a factor is worth one or two LL tests so maybe it's fair that you get more credit? Actually I think I've brought myself through this thought process before and it got nowhere.

I'd like the credit for the GHz days saved with this test

http://www.mersenne.ca/exponent/374011241

The 13,200 for the TF would be nice but the 2X10^19 for the 136 bit composite factor would be even better....

[Gordon]

Originally Posted by Gordon

I'd like the credit for the GHz days saved with this test

http://www.mersenne.ca/exponent/374011241

The 13,200 for the TF would be nice but the 2X10^19 for the 136 bit composite factor would be even better....

Back on this again, have found two more "double factor" (67 bits each) exponents, up in the 800m range, where just the P-1 will take 56,800 GHz days to run.

The LL test will take 22,068 days, same for the DC.

That alone has saved the project from using over 100,000 days of CPU.

Why then is the credit only 0.05?

and yes I did search but couldn't find any rational explanation

[chalsall]

Originally Posted by Gordon

Why then is the credit only 0.05?

You get the credit for the work done, not the work saved.

Otherwise everyone would be TFing up in the 999M range, even though, realistically, we'll never ever do a P-1, let alone a LL, test up there for the reasons you yourself detailed.