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: PR target/41653] Add new function to compute thumb1 code size rtx cost

From: Carrot Wei <carrot at google dot com>
To: ramana dot radhakrishnan at arm dot com
Cc: gcc-patches at gcc dot gnu dot org
Date: Thu, 25 Mar 2010 10:47:38 +0800
Subject: Re: [PATCH: PR target/41653] Add new function to compute thumb1 code size rtx cost
References: <7587b291003240304v756c2b1cgbc6754df3b9c1443@mail.gmail.com> <1269441636.2245.33.camel@e200593-lin.cambridge.arm.com>

Flag -fno-short-enums works. Now the total code size is:
without this patch: 2696068
with this patch:     2692852

And there is no regression for any single application.

thanks
Guozhi

On Wed, Mar 24, 2010 at 10:40 PM, Ramana Radhakrishnan
<ramana.radhakrishnan@arm.com> wrote:
>
> On Wed, 2010-03-24 at 18:04 +0800, Carrot Wei wrote:
>> Hi
>>
>
>> Test:
>> This patch was applied to trunk GCC and tested on qemu without new failure.
>> This patch was also tested with CSiBE without mpeg2(Due to expected enum size
>> is 32, but the actual enum size is 16). With options "-mthumb -Os", the total
>> size is:
>
> Doesn't -fno-short-enums work for you ?
>
>> without this patch: 2655992
>> with this patch: ? ?2652844
>
> Can you make sure that there are no individual surprises in the CSiBe
> runs ?
>
> Ramana
>
>
>

References:
- [PATCH: PR target/41653] Add new function to compute thumb1 code size rtx cost
  - From: Carrot Wei
- Re: [PATCH: PR target/41653] Add new function to compute thumb1 code size rtx cost
  - From: Ramana Radhakrishnan

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