This is the mail archive of the gcc@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: [ARM]Extra load store/instructions compared to gcc-3.4

From: Alexander Monakov <amonakov at ispras dot ru>
To: Alexey Kravets <mr dot kayrick at gmail dot com>
Cc: gcc at gcc dot gnu dot org
Date: Wed, 25 Apr 2012 17:00:40 +0400 (MSK)
Subject: Re: [ARM]Extra load store/instructions compared to gcc-3.4
References: <CAG7KMY6ywHVRXM6jmNNxB7urUNJRJTQ55hRKiz9idt4p5q9GrQ@mail.gmail.com>

Hi,

This is a case when ivopts pass makes too many induction variables, exceeding
the number of available registers.  If you want to debug it, see
ivopts_global_cost_for_size in tree-ssa-loop-ivopts.c and its callers;
perhaps, the issue is that it fails to account for IVs created in inner loops
when processing outer loops.

In any case, I recommend to create a bug report in the GCC bugzilla for this
problem.

Hope that helps,
Alexander

Follow-Ups:
- Re: [ARM]Extra load store/instructions compared to gcc-3.4
  - From: Alexey Kravets

References:
- [ARM]Extra load store/instructions compared to gcc-3.4
  - From: Alexey Kravets

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