Announcing that RB2/RC2/RD2 are drop-in replacements for NXP 8051s is stretching reality quite a bit.
Those sockets are hardly a problem, and can be filled by several vendors.
The real problems are the P89C668, for example. With 64KB flash and 8KB xdata RAM, it is larger than the RD2. Atmel announced a RE2 some time ago, but it was all vapor.
One small silicon company in Texas, TEKMOS, has a real reverse-engineered production P89C668 part, and it works. Unfortunately, the analog Reset circuitry is not identical to the NXP parts, and sometimes reset issues creep up in drop-in applications, for some circuits.
This is a microchip commercial strategy rather than technology.
Currently those NXP 8051 customers are likely being coarced to move on to other NXP MCUs such as LPC11xx. Microchip stepped in to attract those customers by offering SST 8051. But later on, Those poor guys will inevitably be coareced to move on to PIC. We will see.
With NXP, it's all about margins. I would assume that doing a die shrink to bring cost within range of their competitors couldn't be justified, and they prefer to focus their newer processes on 32-bit.
Note that Microchip are not announcing any products here; they are just drawing attention to old products - what they describe as "Legacy (sic) Microcontrollers".
Microchip bought this product line (along with the rest of SST) some years ago.
But I think that NXP are certainly shooting themselves in the foot by abandoning their 8051 range...
Blog Doing Math in FPGAs Tom Burke 2 comments For a recent project, I explored doing "real" (that is, non-integer) math on a Spartan 3 FPGA. FPGAs, by their nature, do integer math. That is, there's no floating-point ...