![]() |
![]() |
#1 |
"J. Gareth Moreton"
Feb 2015
Nomadic
5A16 Posts |
![]()
I recently got some new work from PrimeNet, since the current LL job on one of the worker threads is nearly complete, and I got a heap of ECM tests for Fermat Numbers. However, two of them appear to be identical (although the hash is different). Is this normal or am I overlooking something?
ECM2=[Censored1],1,2,8192,1,260000000,26000000000,3,"2710954639361,2663848877152141313,3603109844542291969,319546020820551643220672513" ECM2=[Censored2],1,2,8192,1,260000000,26000000000,3,"2710954639361,2663848877152141313,3603109844542291969,319546020820551643220672513" The status window in Prime95 reports each one as "F13, ECM 3 curves B1=260000000". Last fiddled with by CuriousKit on 2015-04-09 at 02:58 Reason: typo |
![]() |
![]() |
![]() |
#2 |
Jun 2003
19·283 Posts |
![]()
This is normal. In fact other people running ECM-F would also be getting the exact same assignment lines (with different key).
ECM requires many curves to be run on the target number. The server has chosen "3" as the unit of assignment for tracking purposes. But many such assignments are required to get the ECM done up to a prescribed level. After that, the server will issue new assignments with higher bounds (> 260000000,26000000000). |
![]() |
![]() |
![]() |
#3 |
Romulan Interpreter
"name field"
Jun 2011
Thailand
9,973 Posts |
![]()
Adding to what axn missed to explain, the ECM algorithm is not fully deterministic, it has a "random seed" into it, so two people running the same assignment will look for (slightly) different factors. You are running 3 curves with the first assignment, and another 3 curves with the second. You can change 3 to a higher number and have a higher probability to find a factor, by running more curves, which will take proportionally longer time.
Last fiddled with by LaurV on 2015-04-09 at 03:54 |
![]() |
![]() |
![]() |
#4 |
"J. Gareth Moreton"
Feb 2015
Nomadic
2·32·5 Posts |
![]()
Okay, thanks guys. It just looked a little bit odd at first glance.
|
![]() |
![]() |
![]() |
#5 |
"J. Gareth Moreton"
Feb 2015
Nomadic
2×32×5 Posts |
![]()
One minor question... if I wanted to search for more than 3 curves on a single assignment, I can't just manually change the entry in the worktodo.txt file, can I (the test is currently still queued)? Won't that cause a checksum mismatch?
|
![]() |
![]() |
![]() |
#6 |
"GIMFS"
Sep 2002
Oeiras, Portugal
22·32·43 Posts |
![]()
You can do it.
Stop and exit Prime95, edit the worktodo.txt file to whatever number of curves you wish to run, restart Prime95. Upon completion the result will be reported to the server and accepted without problems. You shall avoid exceeding the number of prescribed curves for the B1 bound in use, though. Last fiddled with by lycorn on 2015-04-10 at 01:11 |
![]() |
![]() |
![]() |
Thread Tools | |
![]() |
||||
Thread | Thread Starter | Forum | Replies | Last Post |
p4 - 1.8 ghz - LL tests? | joblack | Hardware | 12 | 2009-08-26 20:40 |
Identical PenD Dual-Core CPUs, one is 2.25x faster | NBtarheel_33 | Hardware | 5 | 2008-11-12 03:24 |
Exponents with identical digits | GP2 | Marin's Mersenne-aries | 2 | 2007-11-19 19:03 |
Identical Machines Differ | pic_nic | Linux | 7 | 2004-11-03 08:28 |
Another set of 20 P-1 tests | GP2 | Completed Missions | 2 | 2003-09-29 23:33 |