This improves performance for slices with an element size larger
than 32 bytes and removes loading a value from the maxElems
array for smaller element sizes.

name                 old time/op  new time/op  delta
GrowSlice/Byte       41.4ns ± 2%  41.5ns ± 1%    ~     (p=0.366 n=10+9)
GrowSlice/Int16      51.1ns ± 2%  51.0ns ± 2%    ~     (p=0.985 n=10+10)
GrowSlice/Int        64.0ns ± 1%  64.2ns ± 1%    ~     (p=0.180 n=10+10)
GrowSlice/Ptr        90.8ns ± 1%  90.7ns ± 1%    ~     (p=0.858 n=9+10)
GrowSlice/Struct/24   108ns ± 0%   108ns ± 2%    ~     (p=0.488 n=8+9)
GrowSlice/Struct/32   118ns ± 2%   117ns ± 2%    ~     (p=0.327 n=10+10)
GrowSlice/Struct/40   159ns ± 1%   148ns ± 1%  -6.87%  (p=0.000 n=10+9)

Updates #21588

Change-Id: I443b82972d379b1befa791f9ee468b3adc6bb760
Reviewed-on: https://go-review.googlesource.com/c/143798
Run-TryBot: Martin Möhrmann <martisch@uos.de>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>