The following benchmarks are improved by the amounts shown
(Others unaffected beyond the level of noise.)
Also adds a test to confirm non-SIMD implementation still correct,
even when run on SIMD-capable machine

Benchmark                   old            new
BenchmarkAddVV/100-18    66148.08 MB/s 117546.19 MB/s 1.8x
BenchmarkAddVV/1000-18   70168.27 MB/s 133478.96 MB/s 1.9x
BenchmarkAddVV/10000-18  67489.80 MB/s 100010.79 MB/s 1.5x
BenchmarkAddVV/100000-18 54329.99 MB/s  69232.45 MB/s 1.3x
BenchmarkAddVW/100-18     9929.10 MB/s  14841.31 MB/s 1.5x
BenchmarkAddVW/1000-18   10583.31 MB/s  18674.44 MB/s 1.76x
BenchmarkAddVW/10000-18  10521.15 MB/s  17484.10 MB/s 1.66x
BenchmarkAddVW/100000-18 10616.56 MB/s  18084.27 MB/s 1.7x

Change-Id: Ic9234c41a43f6c5e9d0e9377de8b4deeefc428a7
Reviewed-on: https://go-review.googlesource.com/32211Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>