I'm doing a Monte Carlo simulation in Python in which I obtain a set of intensities at certain 2D coordinates and then fit a 2D Gaussian to them. I'm using the scipy.optimize.leastsq function and it all seems to work well except for the following error:
Intel MKL ERROR: Parameter 6 was incorrect on entry to DGELSD.
The problem occurs multiple times in a simulation. I have looked around and understand it is something to do with a bug in Intel's MKL library. I can't seem to find a solution to the problem and so I was looking at an alternative fitting function In could use. If someone does know how to get rid of the problem that would be good also.
Asked
Active
Viewed 996 times
6

blah1234
- 131
- 1
- 1
- 4
-
I've had errors like this when linking with library for wrong integer size (c interface); e.g. mkl_intel_lp64.lib vs mkl_intel_ilp64.lib . https://www.intel.com/content/www/us/en/developer/tools/oneapi/onemkl-link-line-advisor.html#gs.xbacf0 – Elliot Apr 22 '22 at 15:28
1 Answers
-1
You could try Intel's python distribution. It includes a pre-built scipy optimized with MKL.

woodenToaster
- 264
- 1
- 4
- 13