Bug in Nested Disection algorithm in intel's MKL

dwfmarchant / pyMKL

Python wrappers to Intel MKL routines

MIT License

9 stars 17 forks source link

Bug in Nested Disection algorithm in intel's MKL #6

Closed jcapriot closed 7 years ago

jcapriot commented 8 years ago

Pardiso has a bug in it's nested disection reordering algorithm (both serial and parallel).

Until Intel releases an update to 11.3.4, can you set iparm[1] = 0 at https://github.com/dwfmarchant/pyMKL/blob/master/pyMKL/pardisoSolver.py#L131

For details on the bug see [https://software.intel.com/en-us/forums/intel-math-kernel-library/topic/657934]

rowanc1 commented 8 years ago

Can we check the version of MKL and switch it out on the fly?

dwfmarchant commented 8 years ago

Sorry, haven't really been paying attention to this project for the last few weeks. Any idea if this issue is limited to 11.3.3 or is it all versions prior?

jcapriot commented 8 years ago

all good, as far as i understand this was only an issue with version 11.3.3. However i'm not sure how long you want to support 11.3.3. Do you know if anaconda has updated their install to 11.3.4 yet?

rowanc1 commented 7 years ago

I believe this can be closed?

dwfmarchant commented 7 years ago

I think so...