Non-optimal optimization of for loop

[Scott Jones]

I was helping someone on Gitter (https://gitter.im/JuliaLang/julia) with some benchmarks against C++, and ran into this:

julia> function jl_foo()

                         x = 0

                         for i in 1:1000_000_000

                             x += 1

                         end

                         return x

                     end

jl_foo (generic function with 1 method)

julia> @code_native jl_foo()

.section __TEXT,__text,regular,pure_instructions

pushq %rbp

movq %rsp, %rbp

xorl %eax, %eax

movabsq $5868546160, %rcx       ## imm = 0x15DCAE870

movq (%rcx), %rcx

cmpq $0, %rcx

je L28

movq %rcx, %rax

L28:

popq %rbp

retq

I would have expected the native code to be simply:

movabsq $1000000000, %rax

retq

Does anybody have an explanation for why it is loading the constant value from memory, and needlessly checking to see if it is 0 (where it leaves the value of 0 in %rax) before copying it into %rax?

Thanks, Scott

[Kristoffer Carlsson]

On 0.4.3 it just returns the value.

[Scott Jones]

Ah, good to know!  So it is a regression then, and hopefully somebody can file an issue about it.  Thanks for checking v0.4.3 (I should have thought to do that, duh!)