FAST-EXECUTE

Reference

MAE/5

Author

M. Anton Ertl (anton@mips.complang.tuwien.ac.at)

Problem

The stack-effect of EXECUTE usually cannot be determined statically. In the context of an optimizing compiler this creates a significant performance problem: no register allocation can be performed across an EXECUTE or across any word that may call EXECUTE directly or indirectly.

The frequent invocation of EXECUTE in object-oriented programs makes it important to avoid this cost.

Proposal

FAST-EXECUTE core-ext

Interpretation:

Interpretation semantics for this word are undefined.

Compilation: ( u1 u2 u3 u4 -- )

Append the run-time semantics given below to the current definition.

Run-time: ( u1*x u3*r xt -- u2*x u4*r )

Remove xt from the stack and perform the semantics identified by it. Other stack effects are due to the word EXECUTEd. An ambiguous definition exists if xt does not have the stack effect ( u1*x u3*r -- u2*x u4*r )

Typical Use

['] + [ 2 1 0 0 ] FAST-EXECUTE

Remarks

This word does not introduce new functionality. It can be implemented (without the performance-enhancing effect) on standard systems with

: FAST-EXECUTE 2DROP 2DROP POSTPONE EXECUTE ; IMMEDIATE

We can therefore wait safely until we have more experience with this word before adopting it.

Experience

none.

Comments

Michael L. Gassanenko:

I think, the new standard must have a section of "experimental words", as FORTH-83 did, and this proposal can go only to this section.

Peter Knaggs (pjk@bcs.org.uk):

I can see very much why you would want this, especially for optimising compilers. The assume command from my stack algebra has the same effect.

Len Zettel (zettel@acm.org):

ANS rules do not allow a standard to include an experimental word set. Were this otherwise, we probably would have included one.

At least in the last go-around, when there was considerable pressure felt by the committee to keep the standard small and the glossaries short, priority was given to defining functionality that a user could not provide himself - things that for one reason or another *had* to be provided by the system.

I believe this bias still exists. A word like FAST-EXECUTE, easily built using a colon definition, would have to show very widespread common practice to get considered.

---

---