This is the mail archive of the gcc-help@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: Optimizer problem with repeated loops over memory

From: Nick Spence <nick dot spence at freescale dot com>
To: Bernd Jendrissek <bernd dot jendrissek at gmail dot com>
Cc: gcc-help at gcc dot gnu dot org
Date: Thu, 15 May 2008 06:59:09 -0700
Subject: Re: Optimizer problem with repeated loops over memory
References: <4829E872.8070807@freescale.com> <b39469680805150314m753ffa54v4e61b626b80e9df1@mail.gmail.com>

Bernd Jendrissek wrote:

On Tue, May 13, 2008 at 9:13 PM, Nick Spence <nick.spence@freescale.com> wrote:
       addr = 0;
       for (i = 0; i < 16; i++) *(addr++) = 0;
I got a conditional jump by replacing your initialization of addr with:

asm("" : "=g" (addr) : "0" (0));

That should make the NULL-ness of addr invisible to GCC.

Thanks - I ended up loading the value from a DRAM base address register, which effectively hid it from the optimizer.

I still don't like the silent corruption of code by the optimizer, and would like to see an optimizer warning message added.

References:
- Optimizer problem with repeated loops over memory
  - From: Nick Spence
- Re: Optimizer problem with repeated loops over memory
  - From: Bernd Jendrissek

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