simplify inline cost computation, update docs

[aviatesk]

No description provided.

[vtjnash]

Why aren't we using

diff --git a/base/operators.jl b/base/operators.jl
index 82b393e0a9..dcc0c3704a 100644
--- a/base/operators.jl
+++ b/base/operators.jl
@@ -636,7 +636,7 @@ for op in (:+, :*, :&, :|, :xor, :min, :max, :kron)
         # note: these definitions must not cause a dispatch loop when +(a,b) is
         # not defined, and must only try to call 2-argument definitions, so
         # that defining +(a,b) is sufficient for full functionality.
-        ($op)(a, b, c, xs...) = afoldl($op, ($op)(($op)(a,b),c), xs...)
+        @inline ($op)(a, b, c, xs...) = afoldl($op, ($op)(($op)(a,b),c), xs...)
         # a further concern is that it's easy for a type like (Int,Int...)
         # to match many definitions, so we need to keep the number of
         # definitions down to avoid losing type information.

since 7bdf9c6 / f8d2294 (#11274 / #13350) / 86ffe49 (#5011)

[aviatesk]

The only difference seems to be that the definition-site annotation is only respected when the call signature is concrete ?

julia/base/compiler/optimize.jl

Lines 296 to 297 in d6f59fa

	if src.inlineable && isdispatchtuple(specTypes)
	# obey @inline declaration if a dispatch barrier would not help

Well, I don't think it makes much difference in reality (we can't get best performance when there is type instability anyway), and so I'm inclined to simply switch to the annotation.

[aviatesk]

@nanosoldier runbenchmarks("union" || "array", vs="@e3b9290e0c8569a321269c2af3fa5dea39bdf450")

[aviatesk]

@nanosoldier runbenchmarks("union" || "array" || "sort" || "foldl", vs="@e3b9290e0c8569a321269c2af3fa5dea39bdf450")

[nanosoldier]

Your benchmark job has completed - possible performance regressions were detected. A full report can be found here.

[nanosoldier]

Your benchmark job has completed - possible performance regressions were detected. A full report can be found here.

[aviatesk]

@nanosoldier runbenchmarks("union" || "array" || "sort" || "foldl", vs="master")

[aviatesk]

@nanosoldier runbenchmarks("union" || "array" || "sort" || "foldl", vs=":master")

[nanosoldier]

Your benchmark job has completed - possible performance regressions were detected. A full report can be found here.