|[5 earlier articles]|
|Re: When to do inline expansion email@example.com (1993-09-21)|
|Re: When to do inline expansion firstname.lastname@example.org (1993-09-21)|
|Re: When to do inline expansion email@example.com (1993-09-22)|
|Re: When to do inline expansion firstname.lastname@example.org (1993-09-22)|
|Re: When to do inline expansion email@example.com (1993-09-22)|
|Re: When to do inline expansion firstname.lastname@example.org (1993-09-25)|
|Re: When to do inline expansion email@example.com (1993-09-27)|
|From:||firstname.lastname@example.org (Steve Simmons)|
|Organization:||Engineering, CONVEX Computer Corp., Richardson, Tx., USA|
|Date:||Mon, 27 Sep 1993 21:42:43 GMT|
>Cons: Many commercial compilers did lousy jobs of optimization on giant
> procedures. Register allocators seemed to suffer a lot (no Sparc
> compiler I know of will use a window push/pop in the middle of a
> routine to spill registers). Call sites are a way for the
> programmer to tell the compiler: "Spill here".
It depends upon the optimizations performed by the compiler. Convex's
interprocedural compiler (Application Procedural Compiler) produced a 400%
improvement in one "real world" case because
it could chose the best loops for vectorization and parallelization through
loop interchange. This is just anecdotal.
Return to the
Search the comp.compilers archives again.