100m p-1 and tf

aurashift · 2016-04-08, 16:01

i can't remember if we addressed this somewhere else...

Can we add a 100M TF and P-1 job type? I'd like to work on and feed my big iron prepared exponents so they can get into LL immediately.

I know this is kind of a niche workfront and I don't want to take any steam away from the project. Maybe we can add it as a secret unlisted work type?

If this leaves a bad taste in anyone's mouth, no worries. Just a wishlist sort of thing.

chalsall · 2016-04-08, 16:58

Quote:

Originally Posted by aurashift

Can we add a 100M TF and P-1 job type? I'd like to work on and feed my big iron prepared exponents so they can get into LL immediately.

GPU72 has a "100M Digits" option, where we release after TF'ing to at least 78 bits (although 81 is "optimal").

Unfortunately not many people are currently working up there at the moment. Since we're now comfortable ahead of most of the LL'ing wave fronts perhaps it's time for some to go back up there....

aurashift · 2016-04-08, 17:52

I don't have any gpu's though :(

kladner · 2016-04-11, 06:57

Quote:

Originally Posted by chalsall

.......Since we're now comfortable ahead of most of the LL'ing wave fronts perhaps it's time for some to go back up there....

Hmm. I might have a look, if other areas are in good shape.

EDIT: OK. I looked at the "Available" tables, and tested what GPU72 was handing out on "Depth First". "Get Assignments" is offering, and delivering, 76-78bit assignments in the 332.3M range, when there are 75 bit available. I will take assignments to at least 78, possibly 79, but I'd really like to start from the lowest bit-level available.

kladner · 2016-04-11, 08:36

Quote:

Originally Posted by kladner

Hmm. I might have a look, if other areas are in good shape.

EDIT: OK. I looked at the "Available" tables, and tested what GPU72 was handing out on "Depth First". "Get Assignments" is offering, and delivering, 76-78bit assignments in the 332.3M range, when there are 75 bit available. I will take assignments to at least 78, possibly 79, but I'd really like to start from the lowest bit-level available.

DOH! I saw LMH, and did not even try the usual Highest, Lowest, etc., with range settings.

Nevermind.

kladner · 2016-04-12, 14:58

Quote:

Originally Posted by aurashift

I don't have any gpu's though :(

I just turned in a 332.3M, taken to 78. I know this is well short of optimal, but the next three levels would take around 36 50-60 hours to run. I guess I am not that committed, at least at the moment.
Edit 2: For some reason, the Status report is listing 77-78 twice. If someone wants a look, I'll PM the exponent.

Quote:

2016-04-12 ktony NF no factor from 2^77 to 2^78
2016-04-12 ktony NF no factor from 2^76 to 2^77
2016-04-12 ktony NF no factor from 2^75 to 2^76
2016-04-11 ktony NF no factor from 2^77 to 2^78

aurashift · 2016-04-12, 16:29

Quote:

Originally Posted by kladner

I just turned in a 332.3M, taken to 78. I know this is well short of optimal, but the next three levels would take around 36 50-60 hours to run. I guess I am not that committed, at least at the moment.
Edit 2: For some reason, the Status report is listing 77-78 twice. If someone wants a look, I'll PM the exponent.

Thank you. TF to 78 takes me the longest. P-1 my stuff can handle well since it's multi-threaded

.

I've got my first few 100M exponents coming in sometime in the next 28 hours. Then they'll start trickling in.

VictordeHolland · 2016-04-12, 16:59

Quote:

Originally Posted by aurashift

Thank you. TF to 78 takes me the longest. P-1 my stuff can handle well since it's multi-threaded

.

I've got my first few 100M exponents coming in sometime in the next 28 hours. Then they'll start trickling in.

How many assignments and how far TF-ed do you need them?

Prime95 · 2016-04-12, 18:12

Quote:

Originally Posted by aurashift

Thank you. TF to 78 takes me the longest. P-1 my stuff can handle well since it's multi-threaded

.

I've got my first few 100M exponents coming in sometime in the next 28 hours. Then they'll start trickling in.

If you must TF in prime95, use 28.9. It is faster and multithreaded.

aurashift · 2016-04-12, 19:36

Quote:

Originally Posted by Prime95

If you must TF in prime95, use 28.9. It is faster and multithreaded.

VictordeHolland · 2016-04-12, 22:53

I've queued a few 100M TF to 78bits expo's on my GPUs, with a bit of luck they'll be ready in time for your machines to grab once they finish their current assignments.

2016-04-08, 16:01	#1
aurashift Jan 2015 11·23 Posts	100m p-1 and tf i can't remember if we addressed this somewhere else... Can we add a 100M TF and P-1 job type? I'd like to work on and feed my big iron prepared exponents so they can get into LL immediately. I know this is kind of a niche workfront and I don't want to take any steam away from the project. Maybe we can add it as a secret unlisted work type? If this leaves a bad taste in anyone's mouth, no worries. Just a wishlist sort of thing.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
To 100M to 75 Bits	petrw1	GPU to 72	6	2017-04-17 04:03
Is LMH > 100M still alive?	TheMawn	LMH > 100M	25	2016-04-18 16:11
Treading LL P-1 for 100M TF	Aramis Wyler	GPU to 72	300	2014-01-20 04:12
Anyone working in 79.3M - 100M ??	markr	Lone Mersenne Hunters	21	2008-12-21 16:02
100M madness	stars10250	Hardware	8	2008-10-02 15:21

2016-04-08, 17:52	#3
aurashift Jan 2015 11·23 Posts	I don't have any gpu's though :(

2016-04-12, 22:53	#11
VictordeHolland "Victor de Hollander" Aug 2011 the Netherlands 498₁₆ Posts	I've queued a few 100M TF to 78bits expo's on my GPUs, with a bit of luck they'll be ready in time for your machines to grab once they finish their current assignments.