From: no.email@nospam.invalid   
      
   Stephen Pelc  writes:   
   > The current VFX 64 bit systems for x64 allow you to select float packs for   
   >   80x87    8 item internal stack   
   >   hfp87    80x87 external stack   
   >   SSE      external stack   
      
   I guess the next thing is to run that same benchmark with the SSE pack.   
      
   With SSE if you want to do transcendentals, is it usual to use a   
   software library that does the numerics?  It seems easier for the   
   library to use the x87 FPU when it is available.   
      
   Maybe some processors will start supporting IEEE 128 bit floating point   
   someday.  I know that the RISC-V architecture contains instructions for   
   it, but I don't know of any processor hardware that does it.   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)