Build 5 seems to be working on my 233MMX, so it looks like I'm back in the game, so many thanks.

Lets hope that the Prime 95 software doesn't have to go through as many bug fix changes as the BOINC software has!

would it help if prime95 checked for updates at startup and then installed the updates automatically
is that a route worth going down

[QUOTE=henryzz;153737]would it help if prime95 checked for updates at startup and then installed the updates automatically
is that a route worth going down[/QUOTE]

When that was suggested years ago, the majority were against the idea.

[QUOTE=Prime95;153748]When that was suggested years ago, the majority were against the idea.[/QUOTE]

I'm still against the idea...

Luigi

[quote=ET_;153750]I'm still against the idea...

Luigi[/quote]
Even if it is optional to disable?

I think it would be quite helpful for some PC's where you haven't quick access to, ore quite many of them....

[QUOTE=Phantomas;153751]Even if it is optional to disable?

I think it would be quite helpful for some PC's where you haven't quick access to, ore quite many of them....[/QUOTE]

- I think to Primenet v4-Primenet v5 transition...
- Not all versions of Prime95 are indicated to download.
- I got a Pentium III happily crunching with version 23.6 when version 24.14 was slower for P III.
- "Borged" PCs usually have slow dial-up connections.
- Pinging the server for a new version is a loss of crunching time :smile: .
- Pushing a new executable version from the server should be programmed from web interface, not from client. Again, bandwidth problems.
- The algorithm should be ready to recognize service versions and different OSes.
- In some cases the software is run from a batch file.

Luigi

[QUOTE=Prime95;153625]Fixed in 25.8 build 5 - now available for Windows only.

The good news is that it wasn't missing any factors - just not reporting found factors properly.[/QUOTE]

George, would it make sense to supplement (or replace) the factor printout with the factor k (i.e. the index k in factor q = 2*k*p+1 of 2^p-1). That could provide a valuable checksum in such cases, and avoid some of the problems with too-long factors in the primenet logs.

[QUOTE=ewmayer;153774]George, would it make sense to supplement (or replace) the factor printout with the factor k (i.e. the index k in factor q = 2*k*p+1 of 2^p-1). That could provide a valuable checksum in such cases, and avoid some of the problems with too-long factors in the primenet logs.[/QUOTE]

I guess a user of Prime95 would be more excited saying "I got factor x" than computing the factor from k. How many gimpsters would understand it?

I had similar doubts using "Factor" (the original one) and learnt the connection between a Mersenne factor and k only after coding my own applet...

Luigi

[QUOTE=ewmayer;153774]George, would it make sense to supplement (or replace) the factor printout with the factor k .[/QUOTE]

I don't see how that is beneficial. The client verifies any factors found (25.8 build 4 had a bug where escaping out of the 5 minute sleep bypassed this check). Plus, the server verifies every incoming factor.

[QUOTE=Prime95;153748]When that was suggested years ago, the majority were against the idea.[/QUOTE]I thing automagic updating is probably a bad idea, but manual checking and (optional) automatic installation of an updated version on user request could be useful. That is, there could be a [i]Help | Check for updates[/i] menu item that would come back with one of:
> "You already have the latest version (v25.x.x)"
or
> "You are running v25.x.x, there is a newer version available (v25.y.y)
[auto-install] [download & manual install] [visit download webpage] [cancel]".

The autoinstall process may take some work to implement, but it should be relatively trivial to add a check-to-see-if-this-version-is-latest feature into Prime95/PrimeNet, with a link to the download page for the latest version, and I think that would be a useful addition.