Delete a random account?? - Page 2

Uncwilly · 2010-06-23, 12:19

Quote:

Originally Posted by cheesehead

What Uncwilly was referring to, I'm sure, was other things "afoot" within Mersenne Forum and GIMPS.

For example our host has global superpowers, yet does not even appear as a super-mod (color of his nym). His post count is a symbol. Etc.

Rodrigo · 2010-06-24, 08:05

Very funny, guys!

I thought it had to be either a joke or some sort of bug in the software.

And it's REALLY effective humor. You had me going there for a while...

Now that I've been warned, I'll be on the lookout for other fanciful stuff.

Rodrigo (still laughing)

RMAC9.5 · 2010-06-26, 05:08

I have been reading the two threads that you started and agree with the other forum members that you are making good progress in understanding how the Prime95 testing process works and am very glad that you have not given up.
In an effort to help you completely understand this testing process, here is a summary of what the others have been telling you in a completely different format.

The only way to tell if a particular number is a Mersenne prime is to do an LL test.
LL tests on prize winning numbers can take a long, long time. Fast, multi-core PCs can test a particular number faster by having all of the cores work on the same number at the same time. These cores share the intermediate calculation results with each other. Separate PCs can NOT "share" a number because they have no way to share the intermediate calculation results with each other.
TF and P-1 Factoring tests can tell you that a particular number is NOT a prime number by finding a factor of that number.
TF and P-1 factoring tests on a particular number take a small amount of time (in comparison with the LL test) and, consequently, are normally done first.
You are absolutely correct that in order for your slower PCs to help your fast PCs on a particular number, the fast ones would "have to twiddle their bits" while the slower ones finished the TF and P-1 tests.
You are also correct in your understanding that the best way for your slower PCs to help the overall GIMPS progress is to do TF and P-1 work on numbers that other people will probably do LL tests on while other people do TF and P-1 tests on numbers that you fast PCs can do LL tests on.

I, too, share your desire to have my older PCs contribute to the over all GIMPS effort. I have an AMD K6 III+ 350 MHz PC running Windows 98. It does TF factoring to low limits. I also have an Intel PII 400 MHz PC, an AMD K7 600 MHz PC, and an Intel PIII 900 MHz PC which run Windows XP and do trial factoring (during the fall and winter to help heat my house).

Rodrigo · 2010-07-26, 17:29

RMAC9.5,

Thank you very much for the explanation, it really helps me to get back up to speed after coming back from a long absence. I'm sorry that it took so long to reply.

I'm glad to hear that you, too, have some slower machines participating in the project, and I'd be curious to learn more about what you have them doing. When you say that you have one PC doing trial factoring "to low limits," what does than mean?

Also, do you have any of these slower PCs doing anything other than TF, or would that not be really practical?

Thanks again.

Rodrigo

RMAC9.5 · 2010-07-28, 04:05

Welcome back Rodrigo,
I am glad you found my post helpful.
To answer your first question, my PC which is doing trial factoring "to low limits" is doing LMH trial factoring. It is too slow and doesn't run enough hours per day to do anything else in a reasonable amount of time.

As Mdettweiler suggested in your other thread, let the server assign LMH-TF work to your 3 oldest PCs or let them work on the Operation Billion Digits project as ET has suggested.

Rodrigo · 2010-07-28, 18:33

Quote:

Originally Posted by RMAC9.5

Welcome back Rodrigo,
I am glad you found my post helpful.
To answer your first question, my PC which is doing trial factoring "to low limits" is doing LMH trial factoring. It is too slow and doesn't run enough hours per day to do anything else in a reasonable amount of time.

As Mdettweiler suggested in your other thread, let the server assign LMH-TF work to your 3 oldest PCs or let them work on the Operation Billion Digits project as ET has suggested.

RMAC9.5,

Thanks for the welcome! It's good to be back, this is a nice community. Or at least it appears to be...

Your post tied all the essential concepts neatly together, describing how they relate to each other. A great overview.

I'll be looking into LMH and OBD for the older PCs over the next day or two. The consensus seems to be that that's where they could do the most good.

Rodrigo

2010-06-24, 08:05	#13
Rodrigo Jun 2010 Pennsylvania 2×467 Posts	A Really Good One! Very funny, guys! I thought it had to be either a joke or some sort of bug in the software. And it's REALLY effective humor. You had me going there for a while... Now that I've been warned, I'll be on the lookout for other fanciful stuff. Rodrigo (still laughing)

2010-06-26, 05:08	#14
RMAC9.5 Jun 2003 3²×17 Posts	Welcome Rodrigo! I have been reading the two threads that you started and agree with the other forum members that you are making good progress in understanding how the Prime95 testing process works and am very glad that you have not given up. In an effort to help you completely understand this testing process, here is a summary of what the others have been telling you in a completely different format. The only way to tell if a particular number is a Mersenne prime is to do an LL test. LL tests on prize winning numbers can take a long, long time. Fast, multi-core PCs can test a particular number faster by having all of the cores work on the same number at the same time. These cores share the intermediate calculation results with each other. Separate PCs can NOT "share" a number because they have no way to share the intermediate calculation results with each other. TF and P-1 Factoring tests can tell you that a particular number is NOT a prime number by finding a factor of that number. TF and P-1 factoring tests on a particular number take a small amount of time (in comparison with the LL test) and, consequently, are normally done first. You are absolutely correct that in order for your slower PCs to help your fast PCs on a particular number, the fast ones would "have to twiddle their bits" while the slower ones finished the TF and P-1 tests. You are also correct in your understanding that the best way for your slower PCs to help the overall GIMPS progress is to do TF and P-1 work on numbers that other people will probably do LL tests on while other people do TF and P-1 tests on numbers that you fast PCs can do LL tests on. I, too, share your desire to have my older PCs contribute to the over all GIMPS effort. I have an AMD K6 III+ 350 MHz PC running Windows 98. It does TF factoring to low limits. I also have an Intel PII 400 MHz PC, an AMD K7 600 MHz PC, and an Intel PIII 900 MHz PC which run Windows XP and do trial factoring (during the fall and winter to help heat my house).

2010-07-26, 17:29	#15
Rodrigo Jun 2010 Pennsylvania 2×467 Posts	Sorry for the long wait RMAC9.5, Thank you very much for the explanation, it really helps me to get back up to speed after coming back from a long absence. I'm sorry that it took so long to reply. I'm glad to hear that you, too, have some slower machines participating in the project, and I'd be curious to learn more about what you have them doing. When you say that you have one PC doing trial factoring "to low limits," what does than mean? Also, do you have any of these slower PCs doing anything other than TF, or would that not be really practical? Thanks again. Rodrigo

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
random comments, random questions and thread titles made for Google	jasong	Lounge	46	2017-05-09 12:32
How can I permanently deactivate/delete my mersenneforum account?	Arkadiusz	Forum Feedback	19	2016-03-11 22:35
delete a random account? ok?	lfm	Lounge	3	2009-08-24 19:26
About random number (random seed) in Msieve	Greenk12	Factoring	1	2008-11-15 13:56
How to delete primes / account	md5	PrimeNet	5	2003-12-15 13:51

2010-07-28, 04:05	#16
RMAC9.5 Jun 2003 3²·17 Posts	Welcome back Rodrigo, I am glad you found my post helpful. To answer your first question, my PC which is doing trial factoring "to low limits" is doing LMH trial factoring. It is too slow and doesn't run enough hours per day to do anything else in a reasonable amount of time. As Mdettweiler suggested in your other thread, let the server assign LMH-TF work to your 3 oldest PCs or let them work on the Operation Billion Digits project as ET has suggested.