XPost: alt.comp.os.windows-11   
   From: phaywood@alphalink.com.au   
      
   Groovy hepcat Mark Lloyd was jivin' in alt.os.linux on Sun, 23 Feb 2025   
   06:54 am. It's a cool scene! Dig it.   
      
   > On Fri, 21 Feb 2025 19:30:50 +0000, Ant wrote:   
   >   
   >> In alt.comp.os.windows-11 Mark Lloyd  wrote:   
   >>> On Fri, 21 Feb 2025 12:11:53 +0100, Carlos E.R. wrote:   
   >>   
   >>> [snip]   
   >>   
   >>> > Yes, an optional separate coprocesor was used initially. If it was   
   >>> > not inserted, an operation trying to use a coprocesor instruction   
   >>> > would raise an exception, which would be attended and the   
   >>> > instruction emulated in software by the program, which was   
   >>> > significantly slower.   
   >>   
   >>> IIRC, the 286 was the first to have that exception, although there   
   >>> were programs to emulate it on older systems. If I remember right,   
   >>> the turbo C compiler would replace every floating point instruction   
   >>> with a call to a routine that checked for a FPU. If it found one it   
   >>> would replace that call with the FPU instruction and return to a few   
   >>> bytes earlier to run that instruction, otherwise it would emulate   
   >>> it.   
   >>   
   >>> [snip]   
   >>   
   >> What about 386SX? IIRC, it didn't have a coprocessor like DX. I can't   
   >> remember!   
   >   
   > No 386 had a coprocessor. You could add a 387 to the system (I don't   
   > remember if there was a 387SX). IIRC, it was also possible to use a   
   > 287.   
      
     Yes, there was a 387SX. And co-processors (or FPUs, as they're more   
   correctly called) for PC family CPUs go right back to the 8087 to go   
   with the 8086/8088. So a PC/XT could have an FPU too.   
      
   > The difference between 386SX and 386DX was a limited (24-bit instead   
   > of 32-bit) address bus, and a multiplexed (32-bit) data bus.   
      
     For anyone who doesn't know what that means; the 386SX had a 16 bit   
   data bus, the same as the 286. All 32 bit memory access instructions   
   had to "thunk" between the internal 32 bit registers and the external   
   data bus, meaning that a 32 bit memory access had to be done in two 16   
   bit accesses.   
      
   > It was with the 486 that the DX meant it had a coprocessor. There was   
   > also a 486DX2 that ran at a doubled clock speed.   
      
     Right. There is also a 486DX4, which runs at 4 times the clock speed.   
   And the 486SX had no built-in FPU, so there is a 487SX to go with it.   
      
   --   
      
      
   ----- Dig the NEW and IMPROVED news sig!! -----   
      
      
   -------------- Shaggy was here! ---------------   
                 Ain't I'm a dawg!!   
      
   --- SoupGate-DOS v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)