Re: [Chicken-hackers] Add branch prediction for C_demand checks [was: Re: [PATCH] Statically determine if argvector can be reused]

[Arthur Maciel]

I wish we had in medicine 10% gains with so non-invasive and pragmatic interventions. 

+1 for both the static argvectors and C_unlikely demand patches. 

Cheers, 

Arthur 

Em 7 de jan de 2017 11:44, "Peter Bex" <address@hidden> escreveu:

On Tue, Dec 27, 2016 at 06:14:42PM +0100, Peter Bex wrote:
> Attached also are benchmark results for unpatched CHICKEN 5, the
> previous patch which adds static argvector checks and results for
> both this patch and the static argvector.

On IRC, Felix asked about the compilation time of the benchmarks,
so here they are:

Original CHICKEN 5:                           3m11s
With static argvectors:                       3m4s
With static argvectors and C_unlikely demand: 2m51s

So that's an overall gain of about 10% with both patches.
Not spectacular, but certainly worthwhile for such relatively
non-invasive patches, I'd say.

Cheers,
Peter

_______________________________________________
Chicken-hackers mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/chicken-hackers