[RFC/ARM] Prefer branches over conditional execution for Cortex-A15.

Richard Earnshaw rearnsha@arm.com
Mon Jan 9 13:33:00 GMT 2012


On 09/01/12 12:01, Matthew Gretton-Dann wrote:
> The attached patch tunes the costs of conditional execution and branches
> for Cortex-A15 to be the same as that for Cortex-A5.
> 
> This gives an average improvement of over 6% on a popular embedded
> benchmark.
> 
> Note that the tuning parameter structure added for Cortex-A15 is only tuned
> for branch costs, and otherwise takes the generic values.  Tuning for other
> optimisations (notably preload insertion) is still to be done.
> 
> Thanks,
> 
> Matt
> 
> gcc/ChangeLog:
> 
> 2012-01-06  Matthew Gretton-Dann  <matthew.gretton-dann@arm.com>
> 
> 	* config/arm/arm-cores.def (cortex-a15): Use cortex_a15_tune for
> 	tuning parameters.
> 	* config/arm/arm.c (arm_cortex_a15_tune): New static variable.
> 
> 

Technically this is too late for 4.7, but the risk is very low and
Cortex-A15 tuning is new.

So OK, but please check in ASAP.

R.



More information about the Gcc-patches mailing list