RangeQuery benchmark optimizations

[pstibrany]

Couple of small optimizations for RangeQuery benchmark.

$ benchstat orig.txt new.txt 
name                 old time/op    new time/op    delta
RangeQuery100000-4     2.53ms ± 8%    2.28ms ± 4%  -10.00%  (p=0.000 n=10+9)
RangeQuery200000-4     7.23ms ± 8%    5.80ms ± 3%  -19.82%  (p=0.000 n=10+9)
RangeQuery500000-4      2.78s ± 2%     2.29s ± 3%  -17.61%  (p=0.000 n=9+10)
RangeQuery1000000-4     5.54s ± 3%     4.54s ± 3%  -18.07%  (p=0.000 n=10+9)

name                 old alloc/op   new alloc/op   delta
RangeQuery100000-4      290kB ± 3%     254kB ± 1%  -12.53%  (p=0.000 n=9+9)
RangeQuery200000-4      586kB ± 7%     368kB ± 1%  -37.22%  (p=0.000 n=10+9)
RangeQuery500000-4      263MB ± 0%     135MB ± 0%  -48.68%  (p=0.000 n=10+10)
RangeQuery1000000-4     527MB ± 0%     271MB ± 0%  -48.62%  (p=0.000 n=10+10)

name                 old allocs/op  new allocs/op  delta
RangeQuery100000-4      2.95k ± 5%     1.97k ± 0%  -33.07%  (p=0.000 n=9+9)
RangeQuery200000-4      7.64k ± 9%     2.09k ± 0%  -72.65%  (p=0.000 n=10+9)
RangeQuery500000-4      4.14M ± 0%     0.14M ± 0%  -96.72%  (p=0.000 n=10+10)
RangeQuery1000000-4     8.27M ± 0%     0.27M ± 0%  -96.75%  (p=0.000 n=10+10)

Benchmark is also updated to have some duplicate streams, as it didn't have any before.

[cyriltovena]

Lgtm