D 2.061 release

Era Scarecrow rtcvb32 at yahoo.com
Mon Jan 7 10:22:44 PST 2013


On Monday, 7 January 2013 at 17:05:19 UTC, Jonathan M Davis wrote:
> On Monday, January 07, 2013 11:13:55 Russel Winder wrote:
>> LDC is though still generally creating faster executables than 
>> DMD. This is just a gut feeling right now, in a couple of 
>> weeks time I'll see if I can get some actual numbers.
>
> I wouldn't expect that to change anytime soon. Work would have 
> to be put in on specifically make dmd generate faster code, and 
> AFAIK, pretty much all of the compiler work is being put into 
> bug fixing and implementing a few new features (like Win64 
> support). I wouldn't expect improved optimizations to enter 
> into the picture much until the bug count is much lower unless 
> someone takes a liking to spending time improving dmd's code 
> generation.

  If I recall right, one of the larger slowdowns was during large 
struct initialization, as dmd wrote individual instructions to 
fill in the struct. I wouldn't think it would be that hard to fix 
that, then compare speeds, they might be closer to equal than we 
think.


More information about the Digitalmars-d mailing list