|Subject:||Re: [patch] Fix behavior of TER on unrolled loops|
|Date:||Thu, 25 Aug 2005 19:37:13 +0200|
David Edelsohn wrote:
My broader point is that clashing over each and every optimization pass is not productive. Most of the developers working on GCC are following a strategy of light-weight, focussed optimization passes and a middle-end mostly free of target dependencies. This is based on experience from other successful compilers -- not an argument based on authority, but a citation that compiler research projects and compiler development teams have experimented and converged on this method as the most effective.
Just out of curiousity, do these other compilers support as many different targets as GCC does?
|<Prev in Thread]||Current Thread||[Next in Thread>|
|Previous by Date:||Re: 4.2 Project: "@file" support, Joe Buck|
|Next by Date:||Re: [PATCH] Fix IVOPTs problem, Jeffrey A Law|
|Previous by Thread:||Re: [patch] Fix behavior of TER on unrolled loops, Richard Kenner|
|Next by Thread:||Re: [patch] Fix behavior of TER on unrolled loops, David Edelsohn|
|Indexes:||[Date] [Thread] [Top] [All Lists]|