Is it the compute libs and that NE10 & cmsis are now depreciated for CortexA V8 64bit?
For large-scale FFT work we recommend the use of the dedicated FFTW library by Frigo and Johnson. The FFTW library is self-optimizing—it automatically tunes itself for each hardware platform in order to achieve maximum performance. It is available under the GNU GPL.
Yeah sort of as was wondering with a big emphasis on mobile and embedded is there a lite-weight libs avail optimised for arm.PocketFFT does seem to be gathering preference as trades in performance minus the bulk but seems like you are saying just use a general purpose lib with optimisations than ARM performance libs as they are only avail at least for free on certain platforms?I found arm-performance-libraries_22.0.2_Ubuntu-20.04_gcc-11.2.tar in between asking and any answer.They seem slightly limited in distro and quite static with versions and also have the worry of working with closed source blobs, so pocketFFT is looking more likely now I have had a reply so thanks.Also I have had further time to think about implementation and implications.
Thanks Amber69 I spotted https://developer.arm.com/downloads/-/arm-performance-libraries but after your reply and some thought with closed source libs think I may go pocketFFT as trades blows with FFTW also has Neon support but has less bulk.Thanks for the reply as did help with thoughts on implementation and implications.