![]() |
![]() |
#1 |
Dec 2009
Peine, Germany
331 Posts |
![]()
1) I looking for concrete information about influence of memory amount assigned to Prime95 with regard to runtime (Stage2 of P-1 & ECM), limits to B1/B2 and efficiency of high RAM assignments in general. Where can find an overview? What I know from several forum posts is that a) it influences B1/B2 (but how much) and b) count of runs (relatives primes) in stage2 (time saved?). I also read that M(1061) is out of ECM range (cause of RAM?) and the thread "Memory available to P-1". What I'd find great for example is a table with B1/B2 values and GHz days for different ram settings.
2) "Out of Memory": Improvement to the software: Don't allow more than 2048MB of RAM on Win32. Search for the reason of this error was annoying. I also get this error if 2 cores try to enter stage 2 although amount is <= 2048 MB leading to a deadlock. (all with version 25.11) 3) Prime95 should be more configurable with regard to ECM. Option "ECM on small Mersenne numbers" always delivers B1=50000 for the region of 5 million (slow CPU, 2 GB RAM for Prime95). I'd like to "Attack smallest M possible" or "Look only for big factors" (big to be specified). 4) Where can I download the help file? "Help" > "Help file" gets me an error message. 5) I'd like an email report service. It would be cool to get factor found reports or LL test results automatically. Right now, I have the keep polling the website "My Account" > "Results". Feel free to split this thread and thanks for your help in advance. |
![]() |
![]() |
![]() |
#2 | |
Jul 2006
Calgary
1A916 Posts |
![]()
Most of these suggestions come under the category of enhancements not likely to get any serious attention. One point you may have missed is any memory amount over 200 megabytes is almost totally insignificant effect on what factors are found. The default until recently (maybe still) was 8MB.
Remember that the development is unpaid and the main developer does it for fun. He has told us the main part he finds fun is the assembly optimization of the central math calculations (or words to that effect, feel free to correct me if I am off base here). The so called user interface and setup are written in a machine independent way to work equally well/poorly and simply on both MS-Windows and Linux/Unix systems. The menus are basically not fun/interesting to work on. That said, there is still some hope some of your ideas may be given more serious consideration. They are not really bad suggestions I just think they may not be given much priority. I'm not in any way in charge of what does or doesn't get done so you need not accept what I'm saying here. Just trying to relay my observations from another user's point of view for you. Quote:
|
|
![]() |
![]() |
![]() |
#3 |
Aug 2002
Termonfeckin, IE
1010110001012 Posts |
![]()
Look at undoc.txt in the Prime95 install folder.
|
![]() |
![]() |
![]() |
Thread Tools | |
![]() |
||||
Thread | Thread Starter | Forum | Replies | Last Post |
amount of memory when P-1 factoring | wildrabbitt | Hardware | 3 | 2015-03-11 16:41 |
a few suggestions for Prime95 | ixfd64 | Software | 7 | 2010-08-26 19:02 |
Suggestions for a new Prime95 version | joblack | Software | 21 | 2009-01-29 03:10 |
Choosing amount of memory | azhad | Software | 2 | 2004-10-16 16:41 |
amount of memory needed for Stage 2 (factor) | pin | Software | 13 | 2003-03-18 19:54 |