fbncsieve - a new fixed n sieve

[rogue]

Originally Posted by pepi37

Doesnot work for me on Win 7 x64

I have a Windows 10 laptop. I don't have any 64-bit WIndows 7 machines available to me?

Did you try to compile? Has anyone else tried to compile and run on Windows 7?

[kruoli]

Originally Posted by rogue

Is anyone else experiencing this problem?

Not working for me on Windows 7, x64, either. I'm definetly using MinGW64. The DLL thus was named libgcc_s_seh_64-1.dll and after renaming to libgcc_s_seh-1.dll, what your program requires, it will give the same 0xc000007b error (I assume at least). I could not find another file with the name.

[rogue]

Which version of mingw64? Do you have the posix version installed? I have attached the dll from my laptop. I don't know why it has the name you mentioned, but that could be a difference between the releases.

[kruoli]

Originally Posted by rogue

Which version of mingw64?

It seems to be an old version from 2015. But the application still crashes when using your provided DLL. I tried to install a new version, but then could not find that file at all (although some libgcc*.dll-files seems to exist).

Originally Posted by rogue

Do you have the posix version installed?

I'm not sure where to check that.

Originally Posted by rogue

I have attached the dll from my laptop.

This file is much smaller than my one but not working, sorry. I'll quickly try on my Windows 10 machine.

[kruoli]

Trying Windows 10, also with your DLL, gave the same error on both executables.

[kruoli]

When using the built-in Windows 10 bash, I was able to compile successfully after adding -std=gnu++11 to the parameter list. Tomorrow, I'll try to cross compile with some tools (from Linux) and with my own MinGW64. It's quite weird.

My self-compiled versions runs excellent. Thank you!

Addendum:

fbncsieve v1.0.0, a CPU program to find factors of k*b^n+c for fixed b, n, and c
Changing p_max to 42088836.  All remaining terms will be prime.
Sieve started: 1 < p < 42088836 with 1410065409 terms
  p=15485863, 10.69K p/sec, 9525905914 factors found, 101.8M f/sec, 36.8% percentDone. ETA 2017-11-27 23:55
Sieve completed at 42088836.  Writing terms file...
...429528886 terms written to k_b3_n10+1.pfgw
Clock time: 100.42 seconds at 25420 p/sec.  Primes tested 2552668.  Factors found: 9570471115.
Processor time: 100.39 sec. (0.05 sieving) (1.00 cores)

[jasonp]

Originally Posted by pepi37

fbncsieve -k20e9 -K30e9 -sk*3^^10+1 ( try both versions)

Did you put the "-sk*3^^10+1" in quotes? When run from a windows command prompt the ^ operator causes string concatenation unless wrapped up in quotes.

[rogue]

Originally Posted by jasonp

Did you put the "-sk*3^^10+1" in quotes? When run from a windows command prompt the ^ operator causes string concatenation unless wrapped up in quotes.

That must be dependent upon the version of Windows. In Windows 10, I use ^^. I didn't try -s"k*3^10+1". I'll try within a single ^ on Windows 10 and see if I can get it to crash. In the minimum, I could give a meaningful error if that is the case.

[rogue]

Originally Posted by rogue

That must be dependent upon the version of Windows. In Windows 10, I use ^^. I didn't try -s"k*3^10+1". I'll try within a single ^ on Windows 10 and see if I can get it to crash. In the minimum, I could give a meaningful error if that is the case.

This: fbncsieve87 -k1e6 -K2e6 -sk*3^10+1 (note the single ^) correctly detects a problem and outputs "Fatal Error: sequence must be in form k*b^n+c where you specify values for b, n and c".

[axn]

There are a few compile errors and warnings when compiling under linux with -std=g++11 option.

I am posting the updated files with fixes for these.

[rogue]

Originally Posted by axn

There are a few compile errors and warnings when compiling under linux with -std=g++11 option.

I am posting the updated files with fixes for these.

Thanks. I have updated my sources.