Three Cool Things about D

[David Nadlinger via Digitalmars-d-announce]

On Monday, 21 December 2015 at 17:28:51 UTC, Andrei Alexandrescu 
wrote:
> https://www.reddit.com/r/programming/comments/3xq2ul/codedive_2015_talk_three_cool_things_about_d/

By the way, even though I wholeheartedly share your sentiment 
regarding those "functional" examples, I found it to be an 
interesting example for the power of modern optimizers that LDC 
manages to compile this

---
ulong factorial(this n) {
	return n <= 1 ? 1 : n * factorial(n - 1);
}
---

to this:

---
__D4test9factorialFkZm:
     mov eax, 1
     cmp edi, 2
     jb  LBB0_7
     mov ecx, edi
     lea edx, [rdi + 7]
     add edi, -2
     mov eax, 1
     test    dl, 7
     je  LBB0_4
     and edx, 7
     neg edx
     mov eax, 1
     .align  4, 0x90
LBB0_3:
     imul    rax, rcx
     dec rcx
     inc edx
     jne LBB0_3
LBB0_4:
     cmp edi, 7
     jb  LBB0_7
     add rcx, -3
     .align  4, 0x90
LBB0_6:
     lea rdx, [rcx + 3]
     imul    rdx, rax
     lea rax, [rcx + 2]
     lea rsi, [rcx + 1]
     imul    rax, rsi
     imul    rax, rdx
     lea rdx, [rcx - 1]
     imul    rdx, rcx
     lea rsi, [rcx - 2]
     imul    rsi, rdx
     imul    rsi, rax
     lea rax, [rcx - 3]
     lea rdx, [rcx - 4]
     imul    rax, rdx
     imul    rax, rsi
     add rcx, -8
     lea edx, [rcx + 3]
     cmp edx, 1
     ja  LBB0_6
LBB0_7:
     ret
---

Not sure about how it arrives at the crazily unrolled loop, but 
no recursion in sight anymore.

  — David