mprime 25.6 - prime95 compatibility

[abstractius]

I'm on OSX having just moved over files from my Vista prime95 "integrated" 2-core calculations, running mprime on those files heats up the CPU as expected but the process gives me no feedback(prints nothing at all in the terminal), no interactivity (can't get the menus to start/stop etc), memory usage remains around 90MB which is neither my 1GB setting nor some default 8MB. Instead mprime -A1 works as expected, as soon as the initial config and download of new tasks is over I get interactivity and all. Any insights on what is happening (especially where all the CPU cycles are going) or how I can find out what is happening appreciated, obviously I am only trying to continue the prime95 calculations but don't mind getting involved with the debugging if indeed some bug has creeped in.

I am appending prime.txt and there maybe something unhealthy about it as it has been "edited" by many versions of the prime95 GUI.

V24OptionsConverted=1
SendAllFactorData=1
StressTester=0
UsePrimenet=1
Windows95Service=1
DialUp=0
UserID=abstractius
UserPWD=*****
AskedAboutMemory=1
WorkPreference=103
Left=0
Top=10
Right=955
Bottom=714
W0=0 0 745 439 0 -1 -1 -1 -1
W1=0 22 767 461 22 -1 -1 -1 -1
W2=0 102 879 541 134 -1 -1 -4 -23
W3=0 80 816 519 71 -1 -1 -4 -23
V5UserID=
OutputIterations=10000
ResultsFileIterations=999999999
DiskWriteTime=10
NetworkRetryTime=2
NetworkRetryTime2=70
DaysOfWork=5
DaysBetweenCheckins=1
TwoBackupFiles=1
SilentVictory=0

[PrimeNet]
Debug=1
ProxyHost=
ProxyUser=
ProxyPass=
ProxyMask=0

[abstractius]

And I could not edit my post: apparently I only needed mprime -d or mprime -m to get the usual interactivity, quite confusing the default is mute! The remaining mystery is that memory use remains 91MB, prime95 was gobbling up the whole gigabyte as soon as it ran, is memory allocation different in mprime? Is it possible to predict when it will grab the memory? Thanks!

I did the little experiment of setting a 32MB limit to mprime, it has not changed the 91MB that appear allocated.

[S485122]

Memory usage limits are only applicaple during stage 2 of P-1 factoring (see various posts on the forum, the FAQ and the wiki). Memory usage can be high as well during the initial stability testing of the machine and torture testing. I suppose you are now just LL testing and the memory used depends only on the size of the exponents and the version of the program.

Jacob

[cheesehead]

Originally Posted by abstractius

UserPWD=

Folks,

Don't post your passwords.

If you discover that you've done so, change the password(s).

Whenever you copy-and-paste from parameter files, consider each copied line thoughtfully.

[cheesehead]

Let me add to what Jacob explained:

Originally Posted by abstractius

memory usage remains around 90MB which is neither my 1GB setting nor some default 8MB.

Originally Posted by abstractius

The remaining mystery is that memory use remains 91MB, prime95 was gobbling up the whole gigabyte as soon as it ran

< snip >

I did the little experiment of setting a 32MB limit to mprime, it has not changed the 91MB that appear allocated.

When you go to Options -> CPU -> Daytime available memory (and Nighttime available memory), "available memory" actually means: how much additional memory (beyond normal use for TF, LL or Stage 1 P-1 or ECM) is to be considered available to allocate for extra workareas to use only during Stage 2 of P-1 or ECM.

The 90-91MB use you're seeing during L-L or during P-1 stage 1 is the basic amount that Prime95 or mprime uses without allocating any extra workareas for use during stage 2. This basic amount may vary, but it's the amount that Prime95/mprime has to have (which is why you're not asked about it and can't set it) to do its work. The "available memory" is in addition to this.