Re: PATCH: [gcc3.5 improvement branch] Very Simple constant propagation

[Andrew Pinski <pinskia at physics dot uc dot edu>]

On Jan 22, 2004, at 15:06, Caroline Tice wrote:

> Hi Roger,
>
> I've been playing around with your patch and comparing it to mine, 
> with the following
> results.


> Benchmark        My Patch          Your Patch
>
> gzip                6.65            12.74
> vpr                20.83            36.98
> gcc               244.13           399.41
> mcf                 1.75             3.28
> crafty             31.95            61.89
> parser             15.61            31.29


I think I can explain this huge compile time differences between the 
two patches.
One thing is that Caroline's patch can allocate a large amount of 
memory, one int and one
pointer (8 to 12 bytes) per psuedo-register (plus the hard registers 
which I had asked her to remove).

While the other hand Roger's patch uses no memory to cache what the 
register points to but
his can produce better code as REG_EQUAL can contain more than 
CONST_INT/CONST_DOUBLE.



Thanks,
Andrew Pinski