crash and unstable number of solutions

[weibingrong]

input.txt
18
2m11+0m12+0m13+0m21+0m22+0m23+0m31+0m32+0m33+0x+0y+0z-2+2a1m11+a4m21+a5m31;
0m11+2m12+1m13+0m21+0m22+0m23+0m31+0m32+0m33+0x+0y+0z-0+2a1m12+a4m22+a5m32;
0m11+1m12+8m13+0m21+0m22+0m23+0m31+0m32+0m33-6x+0y+0z-0+2a1m13+a4m23+a5m33;
0m11+0m12+0m13+2m21+0m22+0m23+0m31+0m32+0m33+0x+0y+0z-0+2a2m21+a4m11+a6m31;
0m11+0m12+0m13+0m21+2m22+1m23+0m31+0m32+0m33+0x+0y+0z-2+2a2m22+a4m12+a6m32;
0m11+0m12+0m13+0m21+1m22+8m23+0m31+0m32+0m33+0x-6y+0z-1+2a2m23+a4m13+a6m33;
0m11+0m12+0m13+0m21+0m22+0m23+2m31+0m32+0m33+0x+0y+0z-0+2a3m31+a5m11+a6m21;
0m11+0m12+0m13+0m21+0m22+0m23+0m31+2m32+1m33+0x+0y+0z-1+2a3m32+a5m12+a6m22;
0m11+0m12+0m13+0m21+0m22+0m23+0m31+1m32+8m33+0x+0y-6z-2+2a3m33+a5m13+a6m23;
0m11+0m12-6m13+0m21+0m22+0m23+0m31+0m32+0m33+5x+0y+0z-0;
0m11+0m12+0m13+0m21+0m22-6m23+0m31+0m32+0m33+0x+5y+0z-0;
0m11+0m12+0m13+0m21+0m22+0m23+0m31+0m32-6m33+0x+0y+5z+1;
m11^2+m12^2+m13^2-1;
m21^2+m22^2+m23^2-1;
m31^2+m32^2+m33^2-1;
m11m21+m12m22+m13m23;
m11m31+m12m32+m13m33;
m21m31+m22m32+m23*m33;

1 save the above content as input.txt and use phc latest version
2 .\phc.exe -b -068717 input.txt output.txt
3.phc crashed

.\phc.exe -b -t8 -01 input.txt output.txt
it will give 1880 solutions

.\phc.exe -b -t8 -0100 input.txt output.txt
it will give 1876 solutions

Is there any way to provide stable and complete solutions? or provide warnings for unstable or random solutions.

Thanks!

[janverschelde]

Thanks for reporting this issue. The blackbox solver (phc -b) uses default values for tolerances and numerical parameters. Those defaults work well for small and medium sized examples, but no longer for larger problems.

Try to work with phc -u which offers a better path tracker.

[weibingrong]

Hi professor,

I have tried -u option several times. But it requires a lot of mathematics and ultimately I failed.
Could you please give some samples or tips?

By the way, if we want to interoperate with C/C++, how should we do?

Thanks a lot and looking forward to your instructions!

[weibingrong]

Hello, I have to disturb you for a moment because phc -u needs many steps to perform Could you please provide the operation steps for phc -u to ensure that these steps has the same effect with phc -b And then I can try to input different parameters to test result. Best regards and thank you for your powerful phc tool!

…

------------------ 原始邮件 ------------------ 发件人: "janverschelde/PHCpack" ***@***.***>; 发送时间: 2024年5月3日(星期五) 上午6:29 ***@***.***>; ***@***.******@***.***>; 主题: Re: [janverschelde/PHCpack] crash and unstable number of solutions (Issue #72) Thanks for reporting this issue. The blackbox solver (phc -b) uses default values for tolerances and numerical parameters. Those defaults work well for small and medium sized examples, but no longer for larger problems. Try to work with phc -u which offers a better path tracker. — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: ***@***.***>