1) What is the latest CMSIS version 5 support for STM32F7 and where can i find it. 2)Also any other methods and mechanisms to perform optimizations for floating-point operations in the given STM32F7 MCU platform.

[PGama]

Currently our customer implemented intensive mathematical floating-point double-precision algorithms in Matlab and converted them into C code using the Matlab Code Conversion tool. He needs to perform further platform-level optimizations using CMSIS DSP-based APIs for operations like matrix multiplication, addition other math functions square root, etc, at double precision (i..e double variables f64) as the first level of optimization. Can you please share the details related to the latest CMSIS version 5 support for STM32F7 series MCU as currently, the STMCube is supporting an older version of the CMSIS version for the STM32F7 series MCU family?

[TDK]

To install DSP libraries in your project:

https://community.st.com/s/article/configuring-dsp-libraries-on-stm32cubeide

Libraries can be selected in CubeMX within Software Packs -> Select Components (Alt+O).

The CMSIS DSP package can be viewed. Mine shows CMSIS v5.7 and DSP v1.8.0 for a random STM32F7 chip.

Note that only certain STM32F7 chips support double precision FPU.