This is the mail archive of the 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, LRA] Remove REG_DEAD and REG_UNUSED notes.

On 09/24/2013 03:40 PM, Mike Stump wrote:
> On Sep 24, 2013, at 12:23 PM, Steven Bosscher <> wrote:
>> On Tue, Sep 24, 2013 at 5:03 PM, Eric Botcazou wrote:
>>>> This patch removes REG_DEAD and REG_UNUSED notes
>>> DF framework is supposed to do it for you.
>> Unfortunately LRA uses its own DF framework.
I'd say fortunately.  Otherwise, LRA would decrease whole compiler speed
by upto 10%.  I can say it because of the first version of LRA was based
on DF framework.

I guess that was one reason why reload was never moved to DF too. 
LRA/reload is the most live analysis intensive passes (too many changes
on several sub-passes).

LRA uses also REG notes internally and recreates them several time as a
by-product of pseudo live-range analysis used for (reload) pseudo

I'd be glad if we move to DF without compilation time degradation.

I already wrote that DF-design (although it has a nice API which is good
for prototyping) has a bad design with resource usage point of view. 
Keeping a lot of duplicated info by RTL side increases memory footprint
very much, worsens data locality, and slows down the compiler.  As I
remember correctly, GCC steering committing permitted 5% compiler time
degradation as a criterium to include it into the trunk and DF achieved
this with difficulties.

> Is that a bug, or a feature?

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