about 24bit floating point in MCU (9s08dz60)

[darwinying]

We are developing an application in MCU (9s08dz60). We need a support package for 24bit floating-point operations, but that CW only supports 32bit and 64 bit. When we use Microchip mcu, we find the  ieee 24bit float support package and compiler. How can we get 24bit floating-point operations

[vicentegomez]

the only option in CodeWarrior are 32 and 64 bits

Regards

Vicente G

[darwinying]

Thank you for your reply!

CW only supports 32bit and 64bit floating point. Can you tell me if there are other development tools supporting nxp mcu that can support 24bit float operation?

[kef2]

You didn't say what for do you need those 24bits FP numbers?

1) Is your old PIC project using them and you want to support old communications protocol? It is simple to convert. Since HC08 is big endian and PIC low endian, you need to change byte order, as well you need to drop least significant byte sending to PIC. 

"1.23" binary representation:

HC08 first byte ..last byte : 3f 9d 70 a4

PIC first byte .. last byte: 70 9d 3f, truncating a4, or 71 9d 3f rounding to nearest

2) Do you want just faster operation heavily sacrificing precision? Fixed point arithmetic can be even faster