topic Re: best way to investigate difference between debug and release when calling LAPACKE_dposv in Intel® oneAPI Math Kernel Library

best way to investigate difference between debug and release when calling LAPACKE_dposv

Le_Callet__Morgan_M — Tue, 25 Oct 2022 13:46:04 GMT

Hello all.

I am using the function LAPACKE_dposv in my c++code and when checking the return value it is not 0 in debug but it is in release.

Shall i be worried ?

Is there some obvious check i should do in my matrices

I am using the following code structure with my own matrix:

LAPACKE_dposv Example Program in C for Row Major Data Layout (intel.com)

Thanks and best regards

Re: best way to investigate difference between debug and release when calling LAPACKE_dposv

VidyalathaB_Intel — Wed, 26 Oct 2022 09:07:24 GMT

Hi Morgan,

Thanks for reaching out to us.

>>I am using the function LAPACKE_dposv in my c++code and when checking the return value it is not 0 in debug but it is in release.

Could you please share your project file with the sample reproducer code and steps to reproduce (if any) so that we could check it from our end?

Meanwhile, you can check for the details of the return value of the function here at this link

https://www.intel.com/content/www/us/en/develop/documentation/onemkl-developer-reference-c/top/lapack-routines/lapack-linear-equation-routines/lapack-linear-equation-driver-routines/posv.html

This function returns a value info.

If info = 0, the execution is successful.
If info = -i, parameter i had an illegal value.
If info = i, the leading minor of order i (and therefore the matrix A itself) is not positive definite, so the factorization could not be completed, and the solution has not been computed.

>>I am using the following code structure with my own matrix:

Could you please confirm if you have observed the same issue with the example code provided in the link?

A quick check of the example in the link from my end shows that the LAPACKE_dpsov function call returns 0 in Debug mode as well.

Additionally, please let us know the MKL version being used in this case and your OS environment details.

Regards,

Vidya.

Re:best way to investigate difference between debug and release when calling LAPACKE_dposv

VidyalathaB_Intel — Wed, 02 Nov 2022 09:21:32 GMT

Hi Morgan,

As we haven't heard back from you, could you please provide us with an update regarding the issue? Please get back to us with the above-mentioned details if the issue still persists.

Regards,

Vidya.

Re: best way to investigate difference between debug and release when calling LAPACKE_dposv

Le_Callet__Morgan_M — Mon, 07 Nov 2022 13:13:56 GMT

Dear Vidyalatha, many thanks.

I am still checking my matrixes to make sure it does not have any non finite number. Once down i ll update you.

Many thanks.

Morgan

Re: best way to investigate difference between debug and release when calling LAPACKE_dposv

Le_Callet__Morgan_M — Tue, 08 Nov 2022 16:49:19 GMT

Dear Vidyalatha, i have checked and the issue was with my code. This thread can be deleted.

Thanks and kind regards

Morgan

Re:best way to investigate difference between debug and release when calling LAPACKE_dposv

VidyalathaB_Intel — Tue, 08 Nov 2022 17:39:14 GMT

Hi Morgan,

>>i have checked and the issue was with my code...This thread can be deleted.

Thanks for the confirmation.

Please post a new question if you need any additional assistance from Intel as this thread will no longer be monitored.

Regards,

Vidya.

Re: best way to investigate difference between debug and release when calling LAPACKE_dposv

Le_Callet__Morgan_M — Thu, 19 Jan 2023 10:00:46 GMT

Dear Dear Vidyalatha, the error was in my code; this thread can be closed and deleted. Thanks and kind regards.