Re: [patch] PR middle-end/39326 - limit LIM

2013-03-11 Thread Steven Bosscher
On Mon, Mar 11, 2013 at 10:52 AM, Richard Biener wrote: > Given > > + well. Return true if all is well, false if something happened > + that is fatal to the rest of the LIM pass. */ > > -static void > +static bool > gather_mem_refs_stmt (struct loop *loop, gimple stmt) > > and > >FOR_EAC

Re: [patch] PR middle-end/39326 - limit LIM

2013-03-11 Thread Richard Biener
On Sun, Mar 10, 2013 at 2:08 AM, Steven Bosscher wrote: > Hello, > > The attached patch fixes another one of the scalability problems > reported in PR middle-end/39326. > > This problem is that tree loop-invariant code motion explodes on basic > blocks with many memory references. Compile time is

[patch] PR middle-end/39326 - limit LIM

2013-03-09 Thread Steven Bosscher
Hello, The attached patch fixes another one of the scalability problems reported in PR middle-end/39326. This problem is that tree loop-invariant code motion explodes on basic blocks with many memory references. Compile time is quadratic in the number of memory references in the basic block, and