COMPLETE!!!! Thinking out loud about getting under 20M unfactored exponents

[petrw1]

Originally Posted by 9vLYvxWUJ7Dy2Hb

Given that I'm only making minimal progress on this range and someone else (Sid & Andy) is having way more throughput I'll limit myself to everything below 31219961


This means the following range is up for grabs again: 31220003 to 31249993

Thanks I'll take it.
We're only a week or so from done.

[petrw1]

Originally Posted by Zhangrc

107.2M is also complete (1991 unfactored).
Please focus on 108.3M which still has 96 factors to go. (I only TFed to find 3 factors, ETA is 2 years, would anyone like to help?)

My current focus is clearing all ranges from 49.6 down.
With the current effort I'm seeing it is probably less than 2 years from complete.

If necessary the next phase will be working from 108.3M up.
The ultimate goal is to complete all ranges up to 999.9M.

I still believe the current TF/P-1 prescribed prior to PRP will clear all ranges (or almost all).
The only caveat is that there is discussion in another thread about reducing the P-1 bounds to save 1 PRP test rather than the previous need to save 2 LL tests.
This may reduce the P-1 factors found ... time will tell.

However, everyone else is free to work wherever and whenever they want.
(Wow, I used the word 'ever' 4 times in one sentence.)

Thanks for your interest.

[petrw1]

Will start in a few days.

[petrw1]

This range only needs 3 more factors.
And with 3x.xM expected to be completely clear within a week 49.6M will be the only holdout above 29.8M
(At my current upper limit of 108.3M)

I still have not heard from Anton's if he is determined to finish 49.6M himself.

I'll watch for activity over the next week or so then decide if I get involved.

As I indicated above there are 388 more exponents that with some hefty P-1 (at least 2M/60M) should do it.
Volunteers welcome. I could start in about a week.
If you are short of P1 power but have some mighty GPU/TF power and went to go to 77 bits I won't get in your way.
Or, P+1 might be a good option too considering how much P-1 has already been done.

Thanks
Wayne

[Zhangrc]

@Anton Repko: Thanks very much for your help at 108.3M, I've already spotted your factor of M108337699.
Now you can do anything below 108.375M, I'll no longer work on the 108.3M range once the PRP wavefront passes it.
(My next target: maybe 108.9M? only 5 to go, but there are only a handful of untested exponents; and most of the P-1 was stage 1 only...)

[petrw1]

Originally Posted by petrw1

I still have not heard from Anton's if he is determined to finish 49.6M himself.

Awesome....I see Anton chugging through the P1 faster than I could. Thanks and fingers crossed.

[Luminescence]

Whoa, those ranges with >60 factors to go are some serious time sinks

Roughly 70 hours to go and my cards will have tested all stage 1-only exponents to 3M/90M. 23 factors to go.

[petrw1]

Originally Posted by Luminescence

Whoa, those ranges with >60 factors to go are some serious time sinks

Roughly 70 hours to go and my cards will have tested all stage 1-only exponents to 3M/90M. 23 factors to go.

Good on you. It's looking good that you'll finish before too long.

Yeah, most of the easy 2x.xM ranges are gone or claimed.
I've just started there and will be working in the medium to hard ranges.
I'm hopeful someone with GPU power will take the worst of the worst to 76 bits.

By my guesstimate 2x.xM should finish within 8 months.
Thanks for your help.

[lycorn]

Looking at the current twok table I noticed the largest stumbling block we have on our way is the 12M range. I would like to work there for a while, so here´s my query:

What would be the most productive/sensible approach?

i) TFing to 73: that would cost ~7.6 GHz-day per trial, for a reasonable expectation of 1 factor per 100 trials. That amounts to circa 760 GHz-days per factor found. Not free...

ii) Extending the current bounds for P-1: many exponents have been tried to relatively low values of B1, B2; there is a significant number with B1 <= 150,000 and B2 less that 20xB1. That might be an interesting opportunity. Using bounds 50x larger than the smallest ones currently tried, the probality of finding a factor is ~ 5.5%, for a cost of 9Ghz-day per trial. That would be less than 180 GHz-days per factor found. It sounds a lot more appealing than TF.

iii) ECM doesn´t appear to be competitive for exponents this large. Running 280 curves @ B1 =50k, B2=5e6 would cost ~ 164 GHz-days for a relatively low probability of finding a factor, giving the P-1 and TF already done.

iv) P+1 might be an option, but I am not sure about what to expect. It seems to me that one should further explore the P-1 option (read: try higher bounds) before going down this route.

Thoughts? Suggestions?
Thanks in advance.

Note: I have permanent access to a GTX1660Ti for TF, and a i5-7400 (Kaby Lake) with 16 GB of memory. I also run several (free) Colab instances.

[lycorn]

On a second thought, analysing point ii) of my previous post, it is to be noted that 760 GHz-days on a GPU are much easier to achieve than 180 on a CPU. So, after all, TF is not as bad as it looks.

[lycorn]

Well, reckon I was still half asleep when I wrote the posts. TFing to 73 costs around 76 Ghz-days, not 7.6. So yeah, TFing is bloody expensive.

Sorry, I promise next time I´ll have my morning coffee before posting