update README

README.md

@@ -23,7 +23,7 @@ A = StructOfArrays(Complex128, 10, 10)
 or by converting an AbstractArray:
 
 ```julia
-A = convert(StructOfArrays, complex(randn(10), randn(10)))
+A = convert(StructOfArrays, complex.(randn(10), randn(10)))
 ```
 
 Beyond that, there's not much to say. Assignment and indexing works as with
@@ -41,7 +41,7 @@ works best with `isbits` immutables such as `Complex{T}`.
 
 ```julia
 using StructsOfArrays
-regular = complex(randn(1000000), randn(1000000))
+regular = complex.(randn(1000000), randn(1000000))
 soa = convert(StructOfArrays, regular)
 
 function f(x, a)
@@ -52,27 +52,44 @@ function f(x, a)
     s
 end
 
-using Benchmarks
+using BenchmarkTools
 @benchmark f(regular, 0.5+0.5im)
 @benchmark f(soa, 0.5+0.5im)
 ```
 
 The time for `f(regular, 0.5+0.5im)` is:
 
 ```
-Average elapsed time: 1.244 ms
-  95% CI for average: [1.183 ms, 1.305 ms]
-Minimum elapsed time: 1.177 ms
+BenchmarkTools.Trial:
+  memory estimate:  64 bytes
+  allocs estimate:  2
+  --------------
+  minimum time:     811.481 μs (0.00% GC)
+  median time:      827.617 μs (0.00% GC)
+  mean time:        845.552 μs (0.00% GC)
+  maximum time:     2.336 ms (0.00% GC)
+  --------------
+  samples:          5880
+  evals/sample:     1
 ```
 
 and for `f(soa, 0.5+0.5im)`:
 
 ```
-Average elapsed time: 832.198 μs
-  95% CI for average: [726.349 μs, 938.048 μs]
-Minimum elapsed time: 713.730 μs
+BenchmarkTools.Trial:
+  memory estimate:  64 bytes
+  allocs estimate:  2
+  --------------
+  minimum time:     491.044 μs (0.00% GC)
+  median time:      517.429 μs (0.00% GC)
+  mean time:        598.599 μs (0.00% GC)
+  maximum time:     4.634 ms (0.00% GC)
+  --------------
+  samples:          8262
+  evals/sample:     1
 ```
 
 In this case, StructsOfArrays are about 1.5x faster than ordinary arrays.
 Inspection of generated code demonstrates that `f(soa, a)` uses SIMD
-instructions, while `f(regular, a)` does not.
+instructions with a optimal data layout, while `f(regular, a)`
+does use SIMD instructions (on Julia 0.7), but does not use a optimal data layout.