topic Gennady in Intel® oneAPI Math Kernel Library

pardiso iparm(2) parameter

Hailong_X_ — Mon, 27 May 2013 18:56:00 GMT

Dear All

From the talbe

http://software.intel.com/en-us/articles/pardiso-parameter-table#table2

It states that

iparm(2)=0 is the MD algorithm and iparm(2)=2 is metis package.

But from the former thread

http://software.intel.com/en-us/forums/topic/299748

"There exists another built-in reordering scheme so called MMD reordering available through iparm(2)=1."

Sergey suggested that iparm(2)=1 is the MD algorithm and iparm(2) = 0 and 2 is metis.

My experience is same as the former thread suggested.

Is the iparm table from your website wrong?

http://software.intel.com/en-us/forums/topic/299748

Hailong

the description of PARDISO

Gennady_F_Intel — Tue, 28 May 2013 18:57:09 GMT

the description of PARDISO parameters ( http://software.intel.com/en-us/forums/topic/299748) is correct.

in the case of iparm(2) == 1 will be used another built-in reordering scheme so called MMD reordering. We doesn't describe this option because of this reoderening is not validated comprehensively.

Gennady

Hailong_X_ — Tue, 28 May 2013 19:32:40 GMT

Gennady

Thanks for your help.

So the description of the table http://software.intel.com/en-us/articles/pardiso-parameter-table#table2 is wrong?

It states iparm(2)=0 is MD not metis.

Hailong