From: pkk@spth.de   
      
   Am 30.11.25 um 10:05 schrieb Michael S:   
   >   
   > The list of supported targets on SDCC front page:   
   > * Intel MCS51 based microprocessors   
      
   OK, that one is plain 8-bit.   
      
   > * Maxim (formerly Dallas) DS80C390 variants   
      
   I'm not really familiar with this one. One one hand it is an MCS-51   
   derivative, and the instructions mostly operate on 8-bit data, but it   
   also AFAIK has a flat 24-bit address space.   
      
   > * Freescale (formerly Motorola) HC08 based   
      
   8-bit, I'd say.   
      
   > * Zilog Z80 based MCUs   
      
   This one gets complicated. The original Z80 had a 4-bit ALU, but is   
   widely considered 8-bit, and I'd agree. It has an 8-bit data bus, a   
   16-bit address bus. Most instructions operate on 8-bit data, but there   
   are some that operate on 16 bits. But SDCC supports many "Z80 based"   
   MCUs. Some are clearly as 8-bit as the Z80, or even more so. But some   
   have quite rich 16-bit instructions, such as the TLCS-90. The eZ80 even   
   has 24-bit registers and quite some instructions that operate on 24-bit   
   data (when in ADL mode - in Z80 mode they operate on 16-bit data   
   instead). The Rabbits up to 3000A also have quite good support for   
   working with 16-bit data.   
      
   > * Padauk (pdk14, pdk15)   
      
   The Padauk µC have an 8-bit wide RAM. Depending on the variant, the   
   program and const data is stored in a PROM or flash that is 14 or 15   
   bits wide.   
      
   > * STMicroelectronics STM8   
      
   Most instructions operate on 8-bit, a few operate on 16 bits. The RAM is   
   connected to the CPU via an 8-bit data bus. The flash is connected to   
   the CPU via a 32-bit data bus (and in some cases alignment affects   
   timing when reading data or instructions from that flash).   
      
   > * MOS 6502 and WDC 65C02   
      
   I think these are quite clearly 8 bit.   
      
   > Work is in progress:   
   > * Rabbit 4000, 5000, 6000   
      
   These are Z80 derivatives. But they also have quite some instructions   
   that operate on 16 or 32 bits. Their data bus can be configured to 8 or   
   16 bits.   
      
   > * Padauk pdk13   
      
   Has a 13 bits wide PROM, otherwise similar to the pdk14/pdk15 above.   
      
   >  and the f8 and f8l   
      
   8-bit or mixed 8/16 bit.   
      
   > Unmaintained:   
   > * Microchip PIC16 and PIC18   
      
   Probably 8 bit.   
      
   > I know nothing about Rabbit and Padauk. The rest of architectures in   
   > the list are '8-bit processors'.   
      
   While I also often use that term, it is not entirely clear where the   
   line should be drawn. Many '8-bit processors' could reasonably be   
   considered 8/16 bit, 16 bit, 24 bit, 32 bit, or combinations thereof   
   instead. But all SDCC targets can in some way address 8-bit bytes.   
      
   Though for the pdk13, pdk14, pdk15 this means just ignoring the upper   
   5/6/7 bits when reading data from PROM/flash - maybe for pdk15, it would   
   make sense to have some optimization for const _BitInt(N) (N <= 15)   
   directly into that PROM/flash, as long as the address is never taken. If   
   SDCC gets pdk16 support, it would probably emulate 8-bit addressing for   
   the 16-bit PROM/flash.   
      
   > Now, if you ask me, I don't understand why Waldek Hebisch considers   
   > difference between 8-bit and [byte-addressable] 16-bit targets   
   > important. As far as size of relevant C types goes, they look the same:   
   > char - 8 bits   
   > int - 16 bit   
   > long - 32 bits   
   > There is possibly difference in the size of 'short', but I don't   
   > understand why it matters.   
      
   If int is 16 bits, shorts is 16 bits, too.   
      
   Philipp   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)