What is the use of CompactRIO if floating point functions and matrices are not allowed to processed by LABVIEW FPGA?

[vishnu123]

Hi,
I am using CompactRIO and i want to implement some control algorithms which mostly involve matrices. I have red in LABVIEW FPGA manual that in this floating point math are not allowed and also matrices are also not allowed, then what is the use of this hardware for implementing complex control algorithms? My all algorithms involve huge matix calculations.Is there any way to overcome these difficulties to use FPGA?

Regards,
Vishnu.

[GerdW]

Hi Vishnu,

"Is there any way to overcome these difficulties to use FPGA?"

Yes: use better algorithms!

Better here means to use algorithm suited for running on a FPGA target! Remember: a FPGA is limited in processing power/processing "space" (gatters to run your code)! Your algorithms have to fit into a certain number of elementary gatter cells and this involves mathematic functions as well... You may have floating point on a FPGA too, but this will need a notable count of gatters to compute...
Remember: you don't use an all-purpose processor here, you just have a (quite huge) bunch of elementary gatter (boolean) functions!


Message Edited by GerdW on 01-28-2008 10:59 AM

[vishnu123]

hi,
well , i wanted to know weather a floating point arithmetic can be implemented using LABVIEW FPGA in my CompactRIO module.As we recieve data from analog module is of binary value and we have to again calibrate to convert into engineering units which involves multiplication of 1e-9 which is very huge value after that i will be processing that value and i will calculate control effort using the measured value here i needed floating point arithmetic and again i will be sending calculated value to analog output module so before sending it should represent binary value compatable with output channle so,. how could i can done all these in FPGA VI in CompactRIO?

[GerdW]

Hi vishnu,

stick with your "binary" data. Use fixed-point arithmetic. Use integer arithmetic. Don't do any unneccessary conversions.

[griffindore]

There's no support for floating point in LabVIEW FPGA. You can find the Fixed-Point Math Library for LabVIEW FPGA here. There's also an FFT IP that uses the Fixed-Point Math Library found on IPNet. You may also want to consider using the NI Digital Filter Design Toolkit for LabVIEW. There's also a PID VI as well as other control VIs available on the FPGA Functions Palette (these are not fixed-point). I hope this helps!

Message Edited by Support on 05-26-2009 09:14 AM