Davieddy's World

[chalsall]

Originally Posted by davieddy

Can you read?

Yes.

Originally Posted by davieddy

Answer mine first.
What are your estimated completion times based on?

The GHz days done over the last 30 days. (Sorry. Perhaps the small text was too small for you to read.)

Originally Posted by davieddy

I advise you stick to 73 bits until you are comfortably clear of the confused LL assignment front, then do some to 74 as resources allow.

Not going to happen.

[Uncwilly]

Originally Posted by chalsall

The GHz days done over the last 30 days. (Sorry. Perhaps the small text was too small for you to read.)

Not going to happen.

When is the projected date where DC-TF should be effectively finished?

[kladner]

Originally Posted by kladner

The ones I have are pledged to 73 as the tables indicated they should be.

But I take higher exponents to 74, also according to the tables.

EDIT: Also, Mr Eddy seems to forget that this is the combined effort of free agents. Any of us are able to follow whichever approaches we see fit. While I tend to follow chalsall's guidelines, I also have my own private motivations. These include doing some "last bit" factoring, some from the lowest TF available, and some in Uncwilly's rarefied altitudes. As mentioned before, I am also (now) running 8 P-1 instances on 32 GB of RAM. (I recently found myself forced into upgrading my main board and CPU. While I was at it I upped the memory, too.)

[davieddy]

Originally Posted by Aramis Wyler

It's getting close enough that we could probably make a betting pool on when the DC-TF would finish, if it weren't so easy to manipulate. I'd guess June 1.

An objection to DC-TF I have, is that if a factor is found, the poor sod who LL-ed the exponent will forever be deprived of confirmation of his/her non-zero residue.
Also, wielding big iron on DCs deprives potential participants with limited firepower of their most sensible work-type.
If the DC front were 1/3 of the LL front, the chance of finding a prime per day would not be that different from first time testing.

D

PS No doubt this has all flown over your heads as usual

[davieddy]

Originally Posted by chalsall

Yes.

But not my posts.

The GHz days done over the last 30 days. (Sorry. Perhaps the small text was too small for you to read.)

It was, and by GHz days I assume you are including the copious "pissing into the wind" stuff, not just preparing new and imminent LL assignments adequately.

Not going to happen.

Because I recommended it?

D

[bcp19]

Originally Posted by davieddy

An objection to DC-TF I have, is that if a factor is found, the poor sod who LL-ed the exponent will forever be deprived of confirmation of his/her non-zero residue.
Also, wielding big iron on DCs deprives potential participants with limited firepower of their most sensible work-type.
If the DC front were 1/3 of the LL front, the chance of finding a prime per day would not be that different from first time testing.

D

PS No doubt this has all flown over your heads as usual

Um... HUH?!?!?!?

Factor = non prime, non-zero residue = non prime...

The exponent is still cleared.

On your other arguement about the number of assignments we are releasing... Rising exponents = longer LL tests and shorter TF tests. With a 32 day lead this is an optimal time to add the extra bit level. If you are truly concerned about meeting the 'demand' then buy a gpu and help out.

[davieddy]

Originally Posted by bcp19

If you are truly concerned about meeting the 'demand' then buy a gpu and help out.

The only help you lot need is the psychiatric variety.

[firejuggler]

*wooosh* I feel the ban- hammer passing by...
Thor( or Zeus) will strike soon!

[Uncwilly]

Originally Posted by davieddy

Also, wielding big iron on DCs deprives potential participants with limited firepower of their most sensible work-type.

In no wise.

Rather it assists them in their task and makes it even more useful. Your argument is akin to saying that we should not use GPU's to take LL candidates to higher levels of TF, because it "deprives participants with CPU's of their most sensible work-type." The more TF that is done, the better chance that each remaining candidate is prime. Also, by doing TF to the appropriate level (adjusted to suit how many LL's or P-1's it might save) helps to ensure that the DC's are not being done on exponents with 'small' factors.

[davieddy]

Originally Posted by firejuggler

*wooosh* I feel the ban- hammer passing by...
Thor( or Zeus) will strike soon!

Play With Fire

With your username this should be right up your street.

D

[davieddy]

Originally Posted by Uncwilly

In no wise.

Rather it assists them in their task and makes it even more useful. Your argument is akin to saying that we should not use GPU's to take LL candidates to higher levels of TF, because it "deprives participants with CPU's of their most sensible work-type." The more TF that is done, the better chance that each remaining candidate is prime. Also, by doing TF to the appropriate level (adjusted to suit how many LL's or P-1's it might save) helps to ensure that the DC's are not being done on exponents with 'small' factors.

Yep.
If the DC-TF effort was supposed to encourage newcomers to do DCs, it would be laudable, although when a factor is found it reduces the quick DCs available.
Unfortunately their motives are less altruistic:
1) They want to perform the DC themselves.
2) They like to find cheap factors (albeit not as cheap as those in the range you seem obsessed with).
3) They are obsessed with the countdown to milestones.

D