This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: Fw: [PATCH 1/2] BZ#13743: PowerPC - Add a new header forplatform specific functions
- From: Michael Meissner <meissner at linux dot vnet dot ibm dot com>
- To: munroesj at us dot ibm dot com
- Cc: Michael Meissner <meissner at linux dot vnet dot ibm dot com>, Michael R Meissner <mrmeissn at us dot ibm dot com>, "Carlos O'Donell" <carlos at systemhalted dot org>, Ulrich Drepper <drepper at gmail dot com>, David Edelsohn <edelsohn at gnu dot org>, "Joseph S. Myers" <joseph at codesourcery dot com>, libc-alpha at sourceware dot org, Roland McGrath <roland at hack dot frob dot com>, ryan dot arnold at gmail dot com, SteveMunroe/Rochester/IBM at IBMUS dot the-meissners dot org, Thomas Schwinge <thomas at codesourcery dot com>, Tulio Magno Quites Machado Filho <tuliom at linux dot vnet dot ibm dot com>
- Date: Thu, 8 Mar 2012 18:08:48 -0500
- Subject: Re: Fw: [PATCH 1/2] BZ#13743: PowerPC - Add a new header forplatform specific functions
- References: <OF25143771.5D6AAC35-ON852579BB.006EE26B-852579BB.006EE6D2@us.ibm.com><20120308202224.GA12510@ibm-tiger.the-meissners.org><1331247636.4861.177.camel@spokane1.rchland.ibm.com>
On Thu, Mar 08, 2012 at 05:00:36PM -0600, Steven Munroe wrote:
> Michael it seems you are responding specifically to timebase and not
> general question (program priority register and data stream control
> register, etc). Or are you accepting the larger set?
>
> Even with Michael gracious offer there are timing and availability
> issues since any new builtins would be GCC-4.8.
In general, it depends. It is always a cost tradeoff in terms of how much work
is it to add to the compiler, how useful is it for the general programmer
vs. using asm's, whether the compiler could fill pipelines, etc. For example,
the x86 GCC compiler added most/all of the new SSE4.1, SSE4.2, SSE4A, AES,
PCLMUL, AVX, AVX2, BMI, TBM, F16C, and MBI2 etc. instructions as builtins.
Now, if you mean things that are more than one instruction, then it depends. I
think the timebase is an example that is better done at the compiler level,
since it can hide the grunge for dealing with 32-bit or 64-bit systems.
I would think for example, it would be nice to have an include file that
defines fields for DBCST, etc.
--
Michael Meissner, IBM
5 Technology Place Drive, M/S 2757, Westford, MA 01886-3141, USA
meissner@linux.vnet.ibm.com fax +1 (978) 399-6899