11-07-2018 10:42 AM - edited 11-07-2018 10:49 AM
I cannot explain the difference and I don't have RIO to test at the moment. Is your RIO linux or older?
Note that the condition number of your matrix is infinite, so most likely none of you "solutions" are good anyway.
11-08-2018 12:16 AM
Hi,
I am using cRIO 9035 Linux platform.
Do you mean Matrix type?
In Matrix-Test.vi: I am having m*n Input Matrix where m will be dyamic size and n will be 33. Therefore solution vector (X) will be of size 33.
I found Matrix*Vector express vi under FPGA Palette. Is this same as Solve Linear Equations.vi?
http://zone.ni.com/reference/en-XX/help/371599N-01/lvfpga/matrixvector/
Thanks & Regards,
Shree
11-08-2018 02:27 AM
@Shree_Nivas wrote:
I found Matrix*Vector express vi under FPGA Palette. Is this same as Solve Linear Equations.vi?
http://zone.ni.com/reference/en-XX/help/371599N-01/lvfpga/matrixvector/
No, that's the opposite.
(Similar to "Solve linear equations" would be "General linear fit", in fact it gives basically the same result but has more tuning options. Have you tried try to compare the different hardware scenarios with more stable inputs. What does the data represent and what are you trying to get out of all this??
(Note that running your code on an 8 core AMD Ryzen lights up only one of the LEDs, meaning that most result elements are slightly different even here.)