Some GPP's/MCU's/DSP's have integrated floating point units as co-processors (for example the: Cortex M4F).
Are their any devices that implement floating point as part of the core architecture - and not as an integrated coprocessing unit ?
No.
I'm trying to understand whether the only difference between a Floating Point and a Fixed point CPU is an add-on Unit (FPU) - like with the Cortex M4 for example.
I was under that that's the case with all "Floating Point" devices. However, stumbnling upon DSP's such as the C6000 from TI made me think otherwise. https://www.ti.com/product/tms320c6671
It seems like this device is Floating Point "from birth" and doesn't utilize an additional FPU co-processor for floating point calculations.
What do you think?
True floating point CPU's are simply fixed point devices with an integrated FPU or are they a completely different breed of devices with profound architectural differences ?