[LLVMdev] Removing -tailcallopt?
davidterei at gmail.com
Fri Feb 5 19:25:48 PST 2010
I've written a new back-end for the GHC Haskell compiler which is at
the stage of being able to bootstrap GHC itself. I use tailcallopt for
all code compiled and am strongly against removing it.
On 6 February 2010 15:27, Jon Harrop <jon at ffconsultancy.com> wrote:
> On Saturday 06 February 2010 02:42:47 Evan Cheng wrote:
>> On Feb 5, 2010, at 7:19 PM, Jon Harrop wrote:
>> > On Friday 05 February 2010 23:35:15 Evan Cheng wrote:
>> >> Does anyone actually using it?
>> > Yes, many LLVM-based projects rely upon TCO to work correctly.
>> Ok, that's all I need to know.
>> >> I'd prefer to just remove it to clean up the implementation if no one
>> >> has any objections.
>> > Are you saying that you want to remove LLVM's working TCO and replace it
>> > with something that is faster but broken?
>> No, I'd rather have something that's working and helps performance.
> I should also mention that tail call optimization is an optimization in space
> and not time: it typically degrades performance (e.g. 2x slower on .NET).
> If you want to improve the performance of tail calls you could either hack the
> existing TCO implementation to make it generate more efficient code, or
> circumvent it when you spot a special case (e.g. direct tail call to self)
> that can be optimized.
> Dr Jon Harrop, Flying Frog Consultancy Ltd.
> LLVM Developers mailing list
> LLVMdev at cs.uiuc.edu http://llvm.cs.uiuc.edu
More information about the llvm-dev