Re: Generating JIT compilers

> >I Think I Will Go For The Template Approach First, I.E. To
> >Generate A Static Mapping From Source To Target Instructions,
> >Which Is Used At Runtime To Emit The Target Instructions.
> >But Before I Reinvent The Wheel... Is Anybody Aware Of Such
> >Work?
>
> The template approach is used in many different sauces:


It Seems That This Approach Is The Common One. For Reasons Of
Speed Or Ease Of Implementation? The Quality Of The Emitted Code
Doesn'T Compare To The Outcome Of A "Real" JIT Compiler, Does
It?


> 3) A cool approach (but I don't have performance data) is to map from
> source language to C and then parse the executable to obtain good
> assembly templates. See Fabrice Bellard's qemu.


that is kind of what our Yirr-Ma framework does. the emitted target
machine code can be generated very fast, yet the quality is by far
not optimal.


> >At The Moment I Tinker Around With Specifications, And How
> >An Instruction Selector For A JIT Compiler Could Be Generated.
>
> Apart from template based, you can try a very simple maximal munch
> algorithm.


In General I Am Interested In The Question Of "What Can Be Generated
From Specs, Without Writing Source/Target Machine Dependent Code" (Of
Course The Specs Must Be Written :-))
As I Started To Look Around For Previous Work, I Could Not Find Much
In That Particular Area, So I Resorted To This News Group :)


cheers,
jens

Return to the comp.compilers page.
Search the comp.compilers archives again.