mersenneforum.org Manual Assignment: Not Needed?
 Register FAQ Search Today's Posts Mark Forums Read

2022-04-29, 16:12   #1
Magellan3s

Mar 2022
Earth

7×11 Posts
Manual Assignment: Not Needed?

I submitted an assignment I got from the manual assignment form and got the following error:

Quote:
 processing: TF factor 26825833831575863180977 for M900011207 (274-275) [range partially factored] (74.506 bits) Error code: 40, error text: Factoring result for M900011207 was not needed
However, when I click on exponent details:
https://www.mersenne.org/report_expo...0011207&full=1

It tells me it has been trial factored up to 81 on primenet and 85 on GPUZ... However, why was this number assigned to me if a factor should have already been found?

I found a factor for that number which should have supposedly already been found it?

Am I misunderstanding something?

2022-04-29, 17:33   #2
axn

Jun 2003

34×67 Posts

Quote:
 Originally Posted by Magellan3s It tells me it has been trial factored up to 81 on primenet and 85 on GPUZ...
mersenne.org doesn't give this information. mersenne.ca does: https://www.mersenne.ca/exponent/900011207

But those are the limits, not the actual TF already done.

The fact that you were given the assignment for 74-75 mean, there was no record of any higher factoring. And we see nothing on mersenne.org that suggests otherwise.

Most likely what happened is, you might've clicked submit twice (without even realizing), and the first submit recorded the factor, and the second submit resulted in the "not needed" error.

2022-04-29, 17:35   #3
kriesel

"TF79LL86GIMPS96gpu17"
Mar 2017
US midwest

11011101111012 Posts

Quote:
 Originally Posted by Magellan3s I submitted an assignment I got from the manual assignment form and got the following error: However, when I click on exponent details: https://www.mersenne.org/report_exponent/?exp_lo=900011207&full=1 It tells me it has been trial factored up to 81 on primenet and 85 on GPUZ... However, why was this number assigned to me if a factor should have already been found? I found a factor for that number which should have supposedly already been found it? Am I misunderstanding something?
Look again. You are credited with finding and submitting a factor at mersenne.org.
Not sure what you meant when you typed GPUZ (graphics card monitoring utility GPU-Z doesn't fit here; maybe GPUto72?)

https://www.mersenne.ca/exponent/900011207 on James Heinrich's server has also updated. (It normally updates from mersenne.org nightly)

It shows 3 rows near page top, respectively:
actual (completed);
primenet goals (for on CPUs);
GPUto72 goals (for on GPUs).

Mersenne.ca lists the exponent as completed to 74 bits, because your run stopped when it found a factor before completing to 75 bits. That's what the * means in its factor found line on mersenne.org.
Attached Thumbnails

Last fiddled with by kriesel on 2022-04-29 at 17:44

2022-04-29, 23:54   #4
Magellan3s

Mar 2022
Earth

7×11 Posts

Quote:
 Originally Posted by kriesel Look again. You are credited with finding and submitting a factor at mersenne.org. Not sure what you meant when you typed GPUZ (graphics card monitoring utility GPU-Z doesn't fit here; maybe GPUto72?) https://www.mersenne.ca/exponent/900011207 on James Heinrich's server has also updated. (It normally updates from mersenne.org nightly) It shows 3 rows near page top, respectively: actual (completed); primenet goals (for on CPUs); GPUto72 goals (for on GPUs). Mersenne.ca lists the exponent as completed to 74 bits, because your run stopped when it found a factor before completing to 75 bits. That's what the * means in its factor found line on mersenne.org.

Kriesel, yes, sorry, it was a typo on my end, GPU72.

So the limit is what the factor should be trial factored to?

Let's say for example
333,333,991:

https://www.mersenne.ca/exponent/333333991
The mersenne.ca lists the limit to 78 for primenet and I have performed a P-1 test for this digit. At this point, would perform a PRP test?

Last fiddled with by Magellan3s on 2022-04-29 at 23:54

2022-04-30, 01:12   #5
Uncwilly
6809 > 6502

"""""""""""""""""""
Aug 2003
101×103 Posts

2·5·13·83 Posts

Quote:
 Originally Posted by Magellan3s https://www.mersenne.ca/exponent/333333991 The mersenne.ca lists the limit to 78 for primenet and I have performed a P-1 test for this digit. At this point, would perform a PRP test?
It is currently at 78. The traditional value that PrimeNet would desire before handing it out for P-1 and then primality testing is 77. But, based upon the advances in TF on GPUs, GPU72 recommends to take it to 81.

Are you doing TF with a GPU? If so, take it to 81.

2022-04-30, 03:15   #6
Magellan3s

Mar 2022
Earth

7·11 Posts

Quote:
 Originally Posted by Uncwilly It is currently at 78. The traditional value that PrimeNet would desire before handing it out for P-1 and then primality testing is 77. But, based upon the advances in TF on GPUs, GPU72 recommends to take it to 81. Are you doing TF with a GPU? If so, take it to 81.
Yes! I have been trial factoring as much as possible with this GPU. I've actually already took M333333991 to 81.

https://www.mersenne.org/report_expo...3333991&full=1

I'm currently in the middle of taking it to 82 and it will finish 82 in 2.5 hrs.

I did a fast P-1 test on it but I think the bounds weren't optimized for the new 30.8 beta so I am currently redoing the P-1 test on that as well.

Code:
[Work thread Apr 29 19:57] Setting affinity to run worker on CPU core #1
[Work thread Apr 29 19:57] Optimal P-1 factoring of M333333991 using up to 57344MB of memory.
[Work thread Apr 29 19:57] Assuming no factors below 2^81 and 1.3 primality tests saved if a factor is found.
[Work thread Apr 29 19:57] Optimal bounds are B1=1838000, B2=226867000
[Work thread Apr 29 19:57] Chance of finding a factor is an estimated 4.71%
Once the P-1 finishes I'll feel more confident about letting the PRP I started on it finish.

Last fiddled with by Magellan3s on 2022-04-30 at 03:17

2022-04-30, 07:54   #7
ET_
Banned

"Luigi"
Aug 2002
Team Italia

24×3×101 Posts

Quote:
 Originally Posted by Magellan3s Yes! I have been trial factoring as much as possible with this GPU. I've actually already took M333333991 to 81. https://www.mersenne.org/report_expo...3333991&full=1 I'm currently in the middle of taking it to 82 and it will finish 82 in 2.5 hrs. I did a fast P-1 test on it but I think the bounds weren't optimized for the new 30.8 beta so I am currently redoing the P-1 test on that as well. Code: [Work thread Apr 29 19:57] Setting affinity to run worker on CPU core #1 [Work thread Apr 29 19:57] Optimal P-1 factoring of M333333991 using up to 57344MB of memory. [Work thread Apr 29 19:57] Assuming no factors below 2^81 and 1.3 primality tests saved if a factor is found. [Work thread Apr 29 19:57] Optimal bounds are B1=1838000, B2=226867000 [Work thread Apr 29 19:57] Chance of finding a factor is an estimated 4.71% Once the P-1 finishes I'll feel more confident about letting the PRP I started on it finish.
Do you have that much memory available?

2022-04-30, 13:07   #8
Magellan3s

Mar 2022
Earth

7·11 Posts

Quote:
 Originally Posted by ET_ Do you have that much memory available?
Yes! This new PC build has an I9 12900K (avx512) (5.2 Ghz) and 64 GB of Corsair DDR5 Ram. I was actually debating putting in two more sticks of 32 GB DDR5 ram (for a total of 128) when I saw the new beta Mprime utilizes ram differently but at \$500 for more memory, it would have to perform DRAMATICALLY better for me to justify the cost!

Last fiddled with by Magellan3s on 2022-04-30 at 13:08

2022-04-30, 17:22   #9
kriesel

"TF79LL86GIMPS96gpu17"
Mar 2017
US midwest

33·263 Posts

Quote:
 Originally Posted by Magellan3s So the limit is what the factor should be trial factored to? Let's say for example 333,333,991: https://www.mersenne.ca/exponent/333333991 The mersenne.ca lists the limit to 78 for primenet and I have performed a P-1 test for this digit. At this point, would perform a PRP test?
The recommended sequence is, for a given Mersenne number, look up the limits & bounds that are appropriate for that one, on mersenne.ca; TF first to an appropriate limit, which is for 100Mdigit, 81 (or for the really extreme 1/64 SP/DP GPUs like GTX 16xx or RTXxxxx, perhaps 82) bits on GPU; adequate P-1 bounds the first time, which implies either explicit bounds or allocating a lot of memory, or both, depending on the application; PRP with GEC & proof generation to a near optimal proof power, which takes a lot of disk space, & must be allowed for preallocation in mprime/prime95.
This sort of thing is covered in https://www.mersenneforum.org/showpo...7&postcount=12 and somewhat in the background reference post, item 35.

Quote:
 Originally Posted by Magellan3s This new PC build has an I9 12900K (avx512) (5.2 Ghz) and 64 GB of Corsair DDR5 Ram.
It's only a dual-channel memory interface CPU, not quad, so more DIMMs would only help P-1, not PRP. https://ark.intel.com/content/www/us...-5-20-ghz.html
In Mlucas, 64 GiB ram should be enough to do even OBD P-1 stage 2 factoring.

Have you tried mfakto or gpuowl on the UHD770 IGP?

Last fiddled with by kriesel on 2022-04-30 at 17:37

 Similar Threads Thread Thread Starter Forum Replies Last Post Ensigm PrimeNet 3 2020-10-13 13:02 greg PrimeNet 3 2020-05-07 16:58 3mg4 PrimeNet 9 2016-12-13 18:24 Unregistered Information & Answers 12 2010-10-07 16:15 JuanTutors Software 2 2008-12-02 01:57

All times are UTC. The time now is 16:07.

Thu Dec 8 16:07:10 UTC 2022 up 112 days, 13:35, 0 users, load averages: 1.98, 1.41, 1.20