Thinking out loud about getting under 20M unfactored exponents

[petrw1]

https://www.mersenne.ca/status/tf/0/0/1/0

[pinhodecarlos]

So prime95 is majority of the time only allocating 8GB for P-1 Stage 2 when I've seen it using 12-13GB before after I re save the local.txt memory requirements to use the 12/13GB. What am I doing wrong?

[petrw1]

Originally Posted by pinhodecarlos

So prime95 is majority of the time only allocating 8GB for P-1 Stage 2 when I've seen it using 12-13GB before after I re save the local.txt memory requirements to use the 12/13GB. What am I doing wrong?

How many workers do you have doing P-1?
If it is more than one then the total allocated to all workers will not exceed about 95% of what you have specified in local.txt.

8G will still get you at least a third of the Relative Primes for one Stage 2 which will be very efficient.

[pinhodecarlos]

1 worker 4 threads.

[petrw1]

Originally Posted by pinhodecarlos

1 worker 4 threads.

You'd probably see better thruput with 4/1 or 2/2.
When it allocates 8GB does it process ALL relative primes? For example something like:

[Oct 19 11:18] Available memory is 14000MB.
[Oct 19 11:18] Using 8000MB of memory.  Processing 480 relative primes (0 of 480 already processed).

[pinhodecarlos]

Originally Posted by petrw1

You'd probably see better thruput with 4/1 or 2/2.

This is a laptop and worried about cooling, this is a better option for me.

Originally Posted by petrw1

When it allocates 8GB does it process ALL relative primes? For example something like:

[Oct 19 11:18] Available memory is 14000MB.
[Oct 19 11:18] Using 8000MB of memory.  Processing 480 relative primes (0 of 480 already processed).

Yes but now looking closer I can see when it needs to process 960 relative primes it uses 12GB. We are fine anyway....it is the first time I taking more time to look at client's logs. Thank you indeed.

[VBCurtis]

Originally Posted by masser

Awesome job on 4.0M!

I have "finished" the 14.00M to 14.01M range. If you were willing to run 3 ECM curves at the B1=50K level on each of the 200 remaining candidates in that range, that would be helpful. If that's too boring, we could think of something else.

I've started ECM on the suggested range. My first set is 5 curves at B1=250K; I may adjust that later. A curve at this size on the Surface tablet I'm using takes about 25k seconds, so ought to progress at roughly 5 candidates per week. I'll give it a couple weeks to start.

[masser]

Originally Posted by VBCurtis

I've started ECM on the suggested range. My first set is 5 curves at B1=250K; I may adjust that later. A curve at this size on the Surface tablet I'm using takes about 25k seconds, so ought to progress at roughly 5 candidates per week. I'll give it a couple weeks to start.

Thank you; that will be great. If you are interested, I could make some P-1 tasks available too.

[pinhodecarlos]

My progress is slow: 35/100 done, two factors found.

[masser]

Originally Posted by masser

I've worked on the 14.0M range for the last 4 months: 43 factors found, with a mix of TF, P-1, and a little bit of ECM. We now have 200 to go for the under 2000 goal. If everything goes according to plan, I should complete this range in about 2 years.

We've now been factoring the 14.0M range for 7 months: 100 factors have been found, with a mix of TF, P-1 and a little bit of ECM. 143 factors to go for the under 2000 goal, so we are almost halfway.

Some ECM highlights:
M14099401 has a 89.780 bit factor: 1063184393209412095691752279
M14017981 has a 88.926 bit factor: 588003375432513022023853873
M14006803 has a 83.617 bit factor: 14830679176132537433148793
M14005253 has a 77.906 bit factor: 283076393615085922807961

Some P-1 highlights:
M14018573 has a 105.648 bit factor: 63580734348894934266923905132639
M14057863 has a 98.809 bit factor: 555219273599892685265834731087
M14010797 has a 80.792 bit factor: 2093252408084980613260481

One TF highlight:
M14079067 has a 71.198 bit factor: More ECM would have found this, I think

It will become harder to find factors, so I'm happy to report that we have gotten some help lately from others on the forum. Many thanks!

[petrw1]

Originally Posted by pinhodecarlos

My progress is slow: 35/100 done, two factors found.

Your factor rate is good; I expect about 1 in 30 attempts.
If you want to see if LapTop heat is actually impacted you could try 2 workers / 2 cores each. You'll get better thruput and maybe without extra heat.

I say that because in your setup (1 worker 4 cores) all 4 cores are still working hard.

Thanks for any and all contributions