Join PrimeGrid
Returning Participants
Community
Leader Boards
Results
Other
drummers-lowrise
|
Message boards :
Sieving :
Manual Sieve log - 527023! + 1
| Author |
Message |
|---|
|
|
|
I just received the message described in the log. | |
|
|
|
|
I am buying lottery tickets .... I just nailed another factorial prime. | |
|
|
|
|
TheDawgz are confused ... AFAWeK sieving does not/can not find primes.
____________
There's someone in our head but it's not us. | |
|
|
|
|
Dwagz,
Any idea why this note was presented in the command window log while processing my work unit? | |
|
|
|
TheDawgz are confused ... AFAWeK sieving does not/can not find primes.
Confirmative. | |
|
|
|
|
I suspect I found factors. I have uploaded the results file to the server.
This is my first work unit, so please excuse my ignorance. | |
|
|
|
Dwagz,
Any idea why this note was presented in the command window log while processing my work unit?
Not sure what you are seeing, but sieving does NOT find primes - it eliminates "numbers/candidates" as being composite (divisible by more than 1 and it's self).
You are far from ignorant - want to see a math idiot - check out TheDawgz - we have failed college calculus a grand total of 5 times over 30 years.
____________
There's someone in our head but it's not us. | |
|
JimB Honorary cruncher
Send message
Joined: 4 Aug 11
Posts: 920
ID: 107307
Credit: 989,270,184
RAC: 150,909
                     
|
|
Here is the upload in question:
6223167013703 | 653932!-1
6223262792191 | 601532!-1
6223317932969 | 517376!+1
6223887219371 | 527023!+1
6223928209007 | 433348!-1
The number on the left is a factor of the candidate on the right. You've proved that those candidates are composite and as a result they will be removed from the sieve and not be primality tested. In case anyone is wondering, I do doublecheck those factors before granting credit, so I know they're correct. The fpsieve program prints factors found to the screen, which is why you saw it. | |
|
|
|
|
Exactly.
I think this description is correct:
Sieving: You consider a long list of candidates. After running the sieve, you have proved that some of these candidates are non-prime. For the rest of the candidates, it will still be unknown whether they are prime or non-prime.
Primality testing (D): You consider just one candidate. After the test finishes, you have either proved that the candidate is prime, or proven that the candidate is non-prime.
Primality testing (P): You consider just one candidate. After the test finishes, you have either proved that the candidate is probable prime, or proven that the candidate is non-prime. Almost all probable primes are actual primes, but in rare cases (depending on the type of probabilistic primality test in question) a "probable prime" is really non-prime. This could be called a pseudoprime.
Note that non-prime numbers greater than 1 are called composite numbers.
/JeppeSN
| |
|
|
|
|
Thanks everyone for the clarification and understanding! | |
|
|
|
|
Jim,
Can I/we get some feedback as to the "experimental" label beside the GPU selection radio button when choosing a work unit for download? | |
|
JimBHonorary cruncher
Send message
Joined: 4 Aug 11
Posts: 920
ID: 107307
Credit: 989,270,184
RAC: 150,909
|
Jim,
Can I/we get some feedback as to the "experimental" label beside the GPU selection radio button when choosing a work unit for download?
The program is not a polished, completed one. For whatever reason the author apparently stopped working on it. If you interrupt it, you may lose all the work you've done. It doesn't seem to flush the open file buffer to diskwhen you Ctrl-C it. I've been setting my console screen to hundreds of lines so that even if it fails I can at least copy the screen buffer and edit out the status lines. Not that the program has failed, but I have had the occasional need to stop that program and had to plan ahead for that possibility. While the program does appear to checkpoint, it doesn't necessarily seem to use the checkpoint file if you restart it. Plus the checkpoint is useless if the previous factors found weren't written to disk.
It's faster than CPU sieving. On a single core of my i7-3770 a 1G range takes 77215 seconds on a single core (or about 1/4 of that time using all four cores). fsievecl64 takes about 4000 seconds per 1G on my GPU (a first-generation TITAN). It does share the GPU nicely with GFN sieving, the GFN stuff seems to take the same amount of time whether I'm also running fsievecl64.exe or not. Your mileage may vary as all GPUs are different.
In short, don't run it unless you can leave it running continuously until it finishes the desired range. If you have frequent power failures, your GPU is flaky or your computer reboots by itself sometimes, this program is not for you. | |
|
|
|
|
Excellent info and details. I will sign-up for one work unit just for the learning experience once the challenge starting this weekend completes.
I do have temporary power problems and which is aggravated by the dead battery in the laptop. However, this is a trivial fix - just get a new replacement power supply - for those few seconds were electricity drops and then returns. | |
|
Message boards :
Sieving :
Manual Sieve log - 527023! + 1 |
