This is the mail archive of the gcc-patches@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] |
On Wednesday, November 27, 2002, at 03:46 PM, Pop Sébastian wrote:
You can actually eliminate computed gotos, transforming them into switches.On Wed, Nov 27, 2002 at 03:27:10PM -0500, Daniel Berlin wrote:Yep, but I changed more things from then. I have it in a local tree.On Wednesday, November 27, 2002, at 02:46 PM, Pop Sébastian wrote:On Wed, Nov 27, 2002 at 08:56:21AM -0500, Diego Novillo wrote:Out of curiosity, why 'while (24)' in browse_tree()?It was just for fun :). Is 'while (1)' the canonical way to say infinite loop? Or maybe I have to change this into a 'label+goto': this would make a good test case for the jump elimination pass :-) By the way is there any interest in the rewriting of the - jump elimination pass?Yes.- switch to if/binary-search pass?Yes Did you ever include those patches i had sent you to speed it up?
I will send it to gcc-patches once I teach it how to deal with gimple
stmts.
(I don't remember if you sent me the patch that disables computed gotos,
but I think that's easy to rewrite it if it's missing.)
I want to benchmark it so we can determine when to apply it.Yep, this could help improving code quality when used with DCE, (and even when we have a goto loop with an induction variable...)
Index Nav: | [Date Index] [Subject Index] [Author Index] [Thread Index] | |
---|---|---|
Message Nav: | [Date Prev] [Date Next] | [Thread Prev] [Thread Next] |