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: PR 6394

From: law at redhat dot com
To: David Edelsohn <dje at watson dot ibm dot com>
Cc: Geoff Keating <geoffk at redhat dot com>, dave at hiauly1 dot hia dot nrc dot ca, gcc at gcc dot gnu dot org
Date: Tue, 30 Apr 2002 15:06:47 -0600
Subject: Re: PR 6394
Reply-to: law at redhat dot com

 In message <200204302038.QAA24876@makai.watson.ibm.com>, David Edelsohn 
writes:
 > 	It depends how the pseudo is being used throughout its entire
 > lifetime.  While a GPR is cheaper for the address case producing the bad
 > code, the pseudo may be used in an FP situation (possibly multiple times)
 > for which GPR is significantly more expensive.
Correct.  

 > 	Either generating a single pseudo was bad or, if the uses are
 > distinct, the register allocator should have split the lifetime across
 > multiple registers for the time when each register class is cheaper.
Our allocator doesn't know how to do this.
jeff

Follow-Ups:
- Re: PR 6394
  - From: John David Anglin

References:
- Re: PR 6394
  - From: David Edelsohn

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