This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH,RFC] Disallow reordering of x87 insns while scheduling


On Sat, 16 Apr 2005, Roger Sayle wrote:
> The following short but possibly controversial code should improve
> the quality of floating point code that we generate for the x87.
> ...
>
> 2005-04-16  Roger Sayle  <roger@eyesopen.com>
>
> 	* sched-deps.c (sched_analyze_1): On STACK_REGS targets, x87, treat
> 	all writes to any stack register as a read/write dependency on
> 	FIRST_STACK_REG.
> 	(sched_analyze_2): Likewise, for reads from any stack register.


Now that Vladimir Makarov has presented benchmark results showing
these changes improve performance and compile-time in addition
to reducing code size, Richard Henderson has dropped his previous
objections (on IRC), so I've committed this patch to mainline CVS.

Let me know if there are any problems.

Roger
--


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]