Newer milestone thread

[kracker]

I believe we should split off to a new thread. Exciting mini-project...

[ET_]

Originally Posted by flashjh

I'm starting on 2347031. That's where you left off, right? (after the 100 for tonight)

Correct! How may exponents (so that I will update the remainders)?

Luigi

[lycorn]

Originally Posted by LaurV

In fact, the "less classes" version (not to be confused with "lowest expo" version) of mfaktc would be theoretically the best choice: it has to be the fastest (about 10-20% faster!) for so "short time" assignments,

Hmmm... not quite. IIRC, the "Less Classes" variant was recommended for assignments of the type 300/400M+ from 65 to 66, or something of that sort, because those ones yes, they are really short.
Taking, say, 1.1M from to 62 to 63 bits is not a short assignment, so I don´t think it qualifies for the use of that particular version.

[petrw1]

Originally Posted by lycorn

I would like to propose that we collectively engage in an effort to take to 65 bits all exponents that are below that level.
I have identified 3 different situations:
...

A couple concerns:
1. We don't want to steal valuable GPU72 resources on the LL wavefront.
2. ECM in this range probably already found most factors up to 65 bits.
3. And more effecient to find ths rest?

[ET_]

Originally Posted by lycorn

Hmmm... not quite. IIRC, the "Less Classes" variant was recommended for assignments of the type 300/400M+ from 65 to 66, or something of that sort, because those ones yes, they are really short.
Taking, say, 1.1M from to 62 to 63 bits is not a short assignment, so I don´t think it qualifies for the use of that particular version.

Correct.

From the tests I did, the timings of v0.19 with or without MORE_CLASSES are roughly the same, both CPU-limited.

Anyway, I am running 2 instances of mfaktc 0.19 and four other sieving applications during the night, and let the scheduler do its job...

Luigi

[ET_]

Originally Posted by petrw1

A couple concerns:
1. We don't want to steal valuable GPU72 resources on the LL wavefront.
2. ECM in this range probably already found most factors up to 65 bits.
3. And more effecient to find ths rest?

Indeed.

As long as we are just a bunch of maniacs in a miniproject, All we need is a list of exponents and a reservation thread.

Now, who will take the charge of creating and updating the list?

Luigi

[flashjh]

Originally Posted by ET_

Correct! How may exponents (so that I will update the remainders)?

Luigi

I have the rest spread accross my machines. I'll post when they're done

[chalsall]

Originally Posted by petrw1

A couple concerns:
1. We don't want to steal valuable GPU72 resources on the LL wavefront.
2. ECM in this range probably already found most factors up to 65 bits.
3. And more effecient to find ths rest?

I don't disagree... But some people appear to be interested/amused by this kind of work.

In this effor, I will only be a facilitator of the range(s) some within this community wish to work; This is to try to minimize "toes being stepped on".

My week is already almost fully programmed, but I should have some cycles this weekend. This is actually really easy for me to do, so it will be online by EOD Sunday.

Please determine what ranges Make Sense^(TM), or are at least desired, by Saturday noon (Barbados Time, +1 of EST).

[ET_]

Originally Posted by chalsall

I don't disagree... But some people appear to be interested/amused by this kind of work.

In this effor, I will only be a facilitator of the range(s) some within this community wish to work; This is to try to minimize "toes being stepped on".

My week is already almost fully programmed, but I should have some cycles this weekend. This is actually really easy for me to do, so it will be online by EOD Sunday.

Please determine what ranges Make Sense^(TM), or are at least desired, by Saturday noon (Barbados Time, +1 of EST).

Luigi

[lycorn]

Originally Posted by chalsall

In this effor, I will only be a facilitator of the range(s) some within this community wish to work; This is to try to minimize "toes being stepped on".

Precisely. It wil be just an automated version of the "list of exponents and reservation thread" mentioned by ET_. Good one about the bunch of maniacs, Luigi, I´m a proud member!

@petrw1: While it is certainly true that many of the factors <= 65 bits to be found in the <1M ranges were already cleared by ECM, don´t forget we have many exponents factored to only 63 and 64 bits in ranges as high as 8M and 10M, where ECM has been in very little use.
But note that the proposal was to "take exponents to 65 bits", the finding of factors being a side effect. So...
Gosh, I really became a pretty little maniac.

[chalsall]

Originally Posted by ET_

Luigi

Please remember the rule: no bowing.

We're all awesome around here....