memcpy vs slice copy

[Sergey Gromov]

Mon, 16 Mar 2009 11:36:50 -0700, Walter Bright wrote:

> Don wrote:
>> Oh. I didn't see it was only 6 bytes. And the compiler even KNOWS it's 
>> six bytes -- it's in the asm. Blimey. It should just be doing that as a 
>> direct sequence of loads and stores, for anything up to at least 8 bytes.
> 
> The compiler will replace it with a simple mov if it is 1, 2, 4 or 8 bytes.

It didn't, actually.

I've just filed a patch which should fix this issue:

http://d.puremagic.com/issues/show_bug.cgi?id=2750

For instance, the benchmark in the original post were compiling to

L31:            mov     ECX,EDX
                mov     EAX,6
                lea     ESI,010h[ESP]
                mov     ECX,EAX
                mov     EDI,EDX
                rep
                movsb
                add     EDX,6
                add     EBX,6
                cmp     EBX,011E1A300h
                jb      L31

With my patch it's just

L3A:            lea     ESI,0Ch[ESP]
                mov     EDI,EAX
                movsd
                movsb
                movsb
                add     EAX,6
                add     ECX,6
                cmp     ECX,011E1A300h
                jb      L3A

as it probably was meant to be.