🚀 go-pugleaf

In article <1992Oct30.172550.29256@en.ecn.purdue.edu> wscott@en.ecn.purdue.edu (Wayne H Scott) writes:
>Mark Hahn (mchst12@cislabs.pitt.edu) wrote:
>: I've been wondering why stack machines are so out-of-favor.
>: Yes, modern compilers do a creditable job of using the relative abundance
>: of registers that a modern macine provides.  But they DO spill, no?
>: 
>[many good advantages of stacks deleted]
>: 
>: To rephrase: why is resource allocation into a static set,
>: namely a fixed register set, better than stack allocation?
>: 
>: regards, Mark.
>
>I many disadvantage of a stack machine is directly due to the fact that
>all operations work with the same set of arguments.
>
>You can hardly write two instructions that do not have a data dependance.
>Try pipelining a stack machine or a super-scalar.
>Every instruction needs to read the TOS and the instruction before has
>not yet written the TOS.
>
>Wayne
>-- 
>_________________________________________________________________________
>Wayne Scott             |  INTERNET: wscott@ecn.purdue.edu
>Electrical Engineering  |  BITNET:   wscott%ecn.purdue.edu@purccvm
>Purdue University       |  UUCP:     {purdue, pur-ee}!ecn.purdue.edu!wscott

I suppose pipelining can't really stand alone without the help of other
tricks like forwarding and bypassing. These techiniques, however, do not
have to be private to register based machines only, tho the implementation
could be rather different. 

While many other points against stack machine are well addressed, I feel
this thread may be less obvious as it seems. Could any kind soul in the
real industry throm some lights on that?

Z.Z