shuffle vec/array/tuple futures/streams for benchmark

[wishawa]

This PR shuffles the CountdownFutures and CountdownStreams used in the benchmark, exposing the bigger-than-linear growth in runtime as the number of subfutures increase.

$ critcmp after before
group                after                                  before
-----                -----                                  ------
array::join 10       1.43      2.5±0.06µs        ? ?/sec    1.00  1727.5±54.00ns        ? ?/sec
array::join 100      1.75     27.5±0.40µs        ? ?/sec    1.00     15.7±0.33µs        ? ?/sec
array::join 1000     7.99  1423.9±39.19µs        ? ?/sec    1.00    178.3±2.53µs        ? ?/sec
array::merge 10      1.37      3.1±0.10µs        ? ?/sec    1.00      2.3±0.03µs        ? ?/sec
array::merge 100     1.00     53.4±0.94µs        ? ?/sec    1.37     73.0±4.45µs        ? ?/sec
array::merge 1000    1.00      2.7±0.05ms        ? ?/sec    1.57      4.2±0.08ms        ? ?/sec
array::race 10       1.35  1164.3±33.54ns        ? ?/sec    1.00   859.3±14.34ns        ? ?/sec
array::race 100      1.73     13.0±0.29µs        ? ?/sec    1.00      7.6±0.12µs        ? ?/sec
array::race 1000     1.96   152.0±10.51µs        ? ?/sec    1.00     77.6±1.39µs        ? ?/sec
tuple::join 10       1.52      2.5±0.04µs        ? ?/sec    1.00  1647.1±34.54ns        ? ?/sec
tuple::merge 10      1.09      3.2±0.04µs        ? ?/sec    1.00      2.9±0.05µs        ? ?/sec
tuple::race 10       1.33  1174.0±31.44ns        ? ?/sec    1.00   883.2±60.63ns        ? ?/sec
vec::join 10         1.43      2.8±0.02µs        ? ?/sec    1.00  1992.8±56.60ns        ? ?/sec
vec::join 100        2.33     44.4±2.02µs        ? ?/sec    1.00     19.1±0.47µs        ? ?/sec
vec::join 1000       10.26 1922.1±63.21µs        ? ?/sec    1.00   187.4±10.52µs        ? ?/sec
vec::merge 10        1.30      3.7±0.05µs        ? ?/sec    1.00      2.8±0.06µs        ? ?/sec
vec::merge 100       1.00     74.7±2.26µs        ? ?/sec    1.11     83.1±2.05µs        ? ?/sec
vec::merge 1000      1.00      4.4±0.26ms        ? ?/sec    1.39      6.1±0.10ms        ? ?/sec
vec::race 10         1.44  1252.1±40.15ns        ? ?/sec    1.00   872.1±12.69ns        ? ?/sec
vec::race 100        1.56     12.2±0.25µs        ? ?/sec    1.00      7.8±0.43µs        ? ?/sec
vec::race 1000       2.01    134.7±2.76µs        ? ?/sec    1.00     67.0±1.37µs        ? ?/sec

[wishawa]

@yoshuawuyts @matheus-consoli Could this and #118, #119 get reviewed? I want to make a crates.io release of a dependant soon and want to have the best join performance.

[matheus-consoli]

This seems to make sense, thank you!
I'll try to review #118 tomorrow.

[scottmcm]

Is this intentional?

Suggested change

fn assert_receiver_is_total_eq(&self) {}