Re: [Open64-devel] question about some old bugs (3082 and 6186)
Brought to you by:
ributzka,
suneeljain
From: Sun C. <sun...@gm...> - 2009-07-31 21:41:36
|
Just a side note. When I did the big 4 way merge last year, I found that the driver code is so vastly different, and the difference are really not logical, just different preferences by different companies. I opted not to merge in the phases.c. My thinking at that time has been, let individual target tailor phases.c for their own interest. I guess with AMD, the question becomes which phase.c should be in the trunk, not so much as why. Sun On Sat, Aug 1, 2009 at 5:04 AM, David Coakley<dco...@gm...> wrote: > Hi all, > > We're working on improving the driver's handling of '-msse3' and related > options, and found that the following code is causing problems: > > (from osprey/driver/phases.c) > > > #ifdef TARG_X8664 > > // Add a workaround for bug 3082 and bug 6186. > > add_string(args, "-mfpmath=387"); > > #endif > > > > > > The referenced bugs are not in the current bugs.open64.net database. > > > > Can anyone offer some insight on why this workaround was needed? Thanks, > > > > -David Coakley / AMD Open Source Compiler Engineering > > > > ------------------------------------------------------------------------------ > Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day > trial. Simplify your report design, integration and deployment - and focus > on > what you do best, core application coding. Discover what's new with > Crystal Reports now. http://p.sf.net/sfu/bobj-july > _______________________________________________ > Open64-devel mailing list > Ope...@li... > https://lists.sourceforge.net/lists/listinfo/open64-devel > > |