Treading LL P-1 for 100M TF

[Aramis Wyler]

Heh.

Anyway, M332283779 Made it to 78 with no factors found. It started on 79 and should be done about 12 hours from this posting and the next one will start.

M332283821 has 1 hour to go to get to 78. No factors found so far. Should finish 79 in about 14.5 hours.

[davieddy]

Back of envelope calculations:

332M = 66M * 1.26⁷
We accept that it is (just) worthwhile taking a 66M exponent to 75 bits.
So my ROT says it is just worth taking 332M to 82 bits (before embarking on a LL test).
The TF and LL test take 25 times longer for 330M than for 66M. Furthermore, the probability of finding a prime is 5 times smaller.

If it just worthwhile TFing 66M to 75, it is clearly desirable to go to 74.
It is possible with our current firepower to do this for most candidates and keep pace with LL completions. Any left over should be taken to 73.
Although Chris agrees with this (indisputable) observation, he was so disappointed at me being right, that he felt the need to create the "My World" thread.

I agree with BCP19 that dabbling with 100M digits is OK in small doses, as long as it isn't diverting a significant amount of firepower from the main goal of finding another prime.

The effort being devoted to the 60M range is expected to find a prime every 4 years.
If the same effort was transferred to the 300M range, we would expect a prime every 500 years.

Does this make my point vividly enough?

David

[kladner]

Originally Posted by davieddy

Back of envelope calculations:

332M = 66M * 1.26⁷
We accept that it is (just) worthwhile taking a 66M exponent to 75 bits.
So my ROT says it is just worth taking 332M to 82 bits (before embarking on a LL test).
The TF and LL test take 25 times longer for 330M than for 66M. Furthermore, the probability of finding a prime is 5 times smaller.

If it just worthwhile TFing 66M to 75, it is clearly desirable to go to 74.
It is possible with our current firepower to do this for most candidates and keep pace with LL completions. Any left over should be taken to 73.
Although Chris agrees with this (indisputable) observation, he was so disappointed at me being right, that he felt the need to create the "My World" thread.

I agree with BCP19 that dabbling with 100M digits is OK in small doses, as long as it isn't diverting a significant amount of firepower from the main goal of finding another prime.

The effort being devoted to the 60M range is expected to find a prime every 4 years.
If the same effort was transferred to the 300M range, we would expect a prime every 500 years.

Does this make my point vividly enough?

David

Yes. The point being that you are subject to paranoid fantasies and delusions of grandeur.

[Uncwilly]

Chris, could you or another super-mod move all of the 'trading P-1 work for 100M digit work' to a different thread? Thanks.

BTW, Since I have been running P-1 in trade, I have found 4 factors, the most recent a 74.15 bit.

[Aramis Wyler]

Originally Posted by davieddy

Back of envelope calculations:

332M = 66M * 1.26⁷
We accept that it is (just) worthwhile taking a 66M exponent to 75 bits.
So my ROT says it is just worth taking 332M to 82 bits (before embarking on a LL test).

So wow, sounds like the appropriate level of factorization (given current hardware) for a LMH is maybe 82, but certainly 81 would be good. Let me look up some quick numbers here... wow, we have exactly 1 LMH that
has been factored well. Since the completion rate is >1, we are definately behind the curve on this one! Maybe I should step up my efforts!

Except of course, I'm just buying p-1 work with this, and until the cuda p-1 program can get p-1 work done at faster than 1/10 of tf speed this is a actually helping the LL range, not hurting it.

[davieddy]

Originally Posted by kladner

Yes. The point being that you are subject to paranoid fantasies and delusions of grandeur.

Did you just think that sentence sounded good?
Could you give me an indication of any relevance to the post you quoted?

I'm not paranoid: they really have got it infamy.

I hope my self-deprecating humour compensates for my hard-earned aura of grandeur.

D

[chalsall]

Originally Posted by davieddy

Could you give me an indication of any relevance to the post you quoted?

We all understand you're just a troll David. (Your motivations we need not speculate on...)

But perhaps the Daviddy's World thread where you argued that we couldn't do what I suggested we could even though I presented evidence is what I suspect kladner is referring to.

And now, in this thread, you're saying going to 74 was your idea all the time (much like you've claimed GPU72 was your idea all the time)...

At some point David, one has to produce to claim....

[davieddy]

Originally Posted by chalsall

But perhaps the Daviddy's World thread where you argued that we couldn't do what I suggested we could even though I presented evidence is what I suspect kladner is referring to.

And now, in this thread, you're saying going to 74 was your idea all the time (much like you've claimed GPU72 was your idea all the time)...

This so petty, but:
I said that we couldn't TF every expo to 74 and keep pace with the rate of LL completion. This remains true ATM.

I was indeed the prime mover in TF via GPU to 72.
I also suggested that it was done in one assignment, and priority should be given to exponents before they were ever allocated for LL.
I said this so often that you described me as a broken record.

Now all this is happening and the wisdom of it taken for granted by most, you are denying I ever suggested it.

I am accused of suggesting sensible strategy.
You grab exponents and dish them out with insufficient thought as if you were lord god almighty.

David

[chalsall]

Originally Posted by davieddy

You grab exponents and dish them out with insufficient thought as if you were lord god almighty.

Examples?

(And, BTW, I'm agnostic; but I do resonate with Buddhism.)

[kladner]

Originally Posted by Aramis Wyler

So wow, sounds like the appropriate level of factorization (given current hardware) for a LMH is maybe 82, but certainly 81 would be good. Let me look up some quick numbers here... wow, we have exactly 1 LMH that
has been factored well. .....

Was that, by chance, to 81? Only one would have to be my work.

EDIT: I see that the exponent (332411273) was handed out the next day for LL. I don't know if that would include P-1.

I am curious. If I were to TF another 332M to a high level, would it be possible for me to carry the assignment over to P-1? I could afford to dedicate one core to such an effort. Would doing P-1 on 332411273 constitute poaching?

[davieddy]

Originally Posted by Aramis Wyler

So wow, sounds like the appropriate level of factorization (given current hardware) for a LMH is maybe 82, but certainly 81 would be good. Let me look up some quick numbers here... wow, we have exactly 1 LMH that has been factored well. Since the completion rate is >1, we are definately behind the curve on this one! Maybe I should step up my efforts!

Since you could take 25 66M exponents to 74 in the same time as one 332M exponent to 81, I am relieved to hear that there are only a few idiots (mostly people round here) who are squandering their massive iron on this currently futile exercise.

The chance of finding a prime per year by LLing 66M expos is 125 times greater than LLing 332M expos.

WAKE UP BCP, KLADNER, CHALSALL and the rest of you non-believers.
Beware of false prophets.

David