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: [tree-ssa] New regressions as of 2003-11-04


In message <20031110194035.GA14398@atrey.karlin.mff.cuni.cz>, Zdenek Dvorak wri
tes:
 >Hello,
 >
 >> > >  >Huh??? Timings of what?
 >> > >
 >> > > When I wrote remove_useless_stmts_and_vars I actually took the time to
 >> > > measure how it impacted compile-time performance.
 >> >
 >> > you wrote it working on cfg? Otherwise I don't see how the timing is
 >> > relevant.  Anyway, my patch does not change cfg at all and the only
 >> > thing it does is a simple pass over all statements, so I don't see a
 >> > potential for it to be slow.
 >> 
 >> This kind of reasoning is why GCC is slowing down all the time.
 >> 
 >> Yes, a pass over all statements _does_ have a compile time impact.  Especia
 >lly 
 >> if we keep adding new ones by the dozens claiming that they don't impact 
 >> compile time.
 >
 >well, the only thing I tried to argue is that the solution I use cannot
 >be slower than the previous one.  I did't not intend to argue about
 >usefulness of the pass, since I believe that people who put its
 >equivalent there did know what they are doing.
But you're version is doing more work than the old one in particular keeping
the CFG up-to-date.  Now it is likely the case that keeping the CFG up-to-date
isn't a huge cost, but then again, if you haven't measured it, how do we 
actually know what that cost is?

Simply saying it cannot be slower does not make it a fact.

Jeff
Jeff

 >
 >Zdenek



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