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: class_narrowest_mode for MODE_INT

From: Jim Wilson <wilson at specifixinc dot com>
To: qinfeng dot zhang at philips dot com
Cc: gcc at gcc dot gnu dot org
Date: Wed, 05 Nov 2003 18:31:49 -0800
Subject: Re: class_narrowest_mode for MODE_INT
References: <OFAE6D5835.FD1CEDE1-ON48256DD6.00087A6B-48256DD6.00094CD4@diamond.philips.com>

qinfeng.zhang@philips.com wrote:

Is the register for saving the comparison result, BImode?

This is target dependent, and you didn't specify a target. If the target has a 1-bit wide register to hold the result of a compare, then yes you can use BImode for the compare result. IA-64 does this for instance.

The register with BImode isn't assigned to a variable with SImode, is it?

You can't put an SImode variable in a BImode register. It won't fit.

You can move values between SImode and BImode registers if the target has defined patterns for them. It is target dependent how and when you can generate them though. For instance, they might be useful for implementing scc operations. -- Jim Wilson, GNU Tools Support, http://www.SpecifixInc.com

References:
- Re: class_narrowest_mode for MODE_INT
  - From: qinfeng . zhang

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