PFGW 3.2.3 has been Released

[rogue]

I have release PFGW 3.2.3. Here is a list of changes:

• Fix an issue with -t when using an input file, PFGW would not reuse bases, thus re-ordering the file would yiled different results. PFGW would use different bases for each test as it not reuse the same base for different numbers.
• Check for sumout errors during primality tests.
• Modified output to pfgw_err.log to match the output shown in GUI.
• Fixed the GUI to show reason that a test has failed.
• Multiply MAXDIFF by 1000 to help avoid SUMOUT errors when for generic forms or when -a1 or -a2 is used.
• Updated to gwnum 25.13 library.
• Attempt to fix issue that prevents WinPFGW from showing result of the test when in quick mode.

You can d/l Windows and MacIntel buildsfrom the following links.

Windows
MacIntel

I will let you know when a linux build is available.

I am not aware of any other bugs with PFGW at this time. If you find any, please let me know.

Please d/l with Safari or Firefox since IE seems to have problems with these files.

--Mark

[Siemelink]

Originally Posted by rogue

I have released PFGW 3.2.3.


--Mark

Goodie, I'll update my copies. I don't know any bugs at the moment, but I have some thoughts.
- I can pass an expression to the PRP test. I would like that expression to show in the log files: PRP 2^7-1 shows up as 127. For huge numbers the expression is more practical.
- Is there a switch that surpresses the output to pfgw.log and pfgw-prime.log? In my scripts I do such output separately.
- the input file must use spaces and can not handle tabs. Could tabs be added, they make the input file easier on the eye.
- In windows, when you hover the mouse over the icon you see: Winpfgw (Running). I already know this because of the colour of the icon. More fun would be if the currently tested string could be shown there: 255*13^123456-1 (33%)
- When in verbose mode, this string shows up all the time: Special modular reduction using zero-padded FFT length 3072 on xxx. I have no use for it with regards to calculations. Is this debug information that can be hidden?

Regards, Willem.

[rogue]

Originally Posted by Siemelink

- I can pass an expression to the PRP test. I would like that expression to show in the log files: PRP 2^7-1 shows up as 127. For huge numbers the expression is more practical.
- Is there a switch that surpresses the output to pfgw.log and pfgw-prime.log? In my scripts I do such output separately.
- the input file must use spaces and can not handle tabs. Could tabs be added, they make the input file easier on the eye.
- In windows, when you hover the mouse over the icon you see: Winpfgw (Running). I already know this because of the colour of the icon. More fun would be if the currently tested string could be shown there: 255*13^123456-1 (33%)
- When in verbose mode, this string shows up all the time: Special modular reduction using zero-padded FFT length 3072 on xxx. I have no use for it with regards to calculations. Is this debug information that can be hidden?

I could probably add a switch that would put both values into the log file.
Why would you suppress output to the log files?
I can investigate the support for tabs, but make no promises.
I added that feature already, but you have to be far enough into a test to see it (at least 2500 iterations).
It is not debug information, thus cannot be hidden.

--Mark

[mdettweiler]

Originally Posted by Siemelink

Is there a switch that surpresses the output to pfgw.log and pfgw-prime.log? In my scripts I do such output separately.

Have you tried just leaving off the -l switch on the command line? That, in my experience, switches off the logging without fail.

[Siemelink]

Originally Posted by Siemelink

- Is there a switch that surpresses the output to pfgw.log and pfgw-prime.log? In my scripts I do such output separately.

Originally Posted by rogue

Why would you suppress output to the log files?

In my script to deliver the k's for a Riesel conjecture I do an extra test when the k is a multiple of the base. This avoids that I have to filter the k later. But that also means that the pfgw.log gets double entries. To have a correct output I do this to a separate file.

Willem.

[rogue]

Originally Posted by Siemelink

In my script to deliver the k's for a Riesel conjecture I do an extra test when the k is a multiple of the base. This avoids that I have to filter the k later. But that also means that the pfgw.log gets double entries. To have a correct output I do this to a separate file.

Willem.

I doubt any change I make will be helpful for what you want to do and if you already have a script to do it, then any effort on my part is unnecessary.

[rogue]

PFGW 3.2.3 for Linux

[henryzz]

Is there any chance of ever having 64-bit builds? Lucas-Lehmer tests can have a 10% speed increase so maybe pfgw can?

[rogue]

Originally Posted by henryzz

Is there any chance of ever having 64-bit builds? Lucas-Lehmer tests can have a 10% speed increase so maybe pfgw can?

Where do you see any claims that LL tests can be 10% faster?

Chances are slim. I don't think that I can build Win64. I can build Mac 64-bit, but there is some ASM that might need to be changed. I can't speak about Linux 64-bit as I don't provide the Linux build. It would have the same potential issues as Mac 64-bit.

[Flatlander]

The "Line in file:" box is blank in WinPFGW3.23.exe

[henryzz]

Originally Posted by rogue

Where do you see any claims that LL tests can be 10% faster?

Chances are slim. I don't think that I can build Win64. I can build Mac 64-bit, but there is some ASM that might need to be changed. I can't speak about Linux 64-bit as I don't provide the Linux build. It would have the same potential issues as Mac 64-bit.

In Prime95 v25.9 the 64-bit version started using the extra registers which came with 64-bit. This gave some improvement over the 32-bit version.