[Draft] Add channel benchmarks

[murfel]

No description provided.

[murfel]

It does perform unexpectedly badly, could you all take a look for any silly mistakes?

The simplest benchmark, sending X elements only (not receiving anything), is already pretty sad:

ChannelBenchmark.sendUnlimited                1000  avgt   10  0.008 ±  0.001  ms/op
ChannelBenchmark.sendUnlimited               10000  avgt   10  0.080 ±  0.001  ms/op
ChannelBenchmark.sendUnlimited              100000  avgt   10  0.800 ±  0.005  ms/op

Full output for the first three counts (4KB, 40KB, 400KB) of Ints (just FYI, no reason to look at this, since the snippet above is already bad enough)

Benchmark                                  (count)  Mode  Cnt  Score    Error  Units
ChannelBenchmark.manySendersManyReceivers     1000  avgt   10  0.119 ±  0.001  ms/op
ChannelBenchmark.manySendersManyReceivers    10000  avgt   10  0.900 ±  0.006  ms/op
ChannelBenchmark.manySendersManyReceivers   100000  avgt   10  9.244 ±  0.418  ms/op
ChannelBenchmark.manySendersOneReceiver       1000  avgt   10  0.108 ±  0.001  ms/op
ChannelBenchmark.manySendersOneReceiver      10000  avgt   10  0.713 ±  0.042  ms/op
ChannelBenchmark.manySendersOneReceiver     100000  avgt   10  7.010 ±  0.115  ms/op
ChannelBenchmark.oneSenderManyReceivers       1000  avgt   10  0.138 ±  0.001  ms/op
ChannelBenchmark.oneSenderManyReceivers      10000  avgt   10  0.923 ±  0.003  ms/op
ChannelBenchmark.oneSenderManyReceivers     100000  avgt   10  8.411 ±  0.035  ms/op
ChannelBenchmark.sendConflated                1000  avgt   10  0.020 ±  0.001  ms/op
ChannelBenchmark.sendConflated               10000  avgt   10  0.187 ±  0.007  ms/op
ChannelBenchmark.sendConflated              100000  avgt   10  1.834 ±  0.013  ms/op
ChannelBenchmark.sendReceiveConflated         1000  avgt   10  0.039 ±  0.001  ms/op
ChannelBenchmark.sendReceiveConflated        10000  avgt   10  0.236 ±  0.009  ms/op
ChannelBenchmark.sendReceiveConflated       100000  avgt   10  1.906 ±  0.019  ms/op
ChannelBenchmark.sendReceiveRendezvous        1000  avgt   10  0.103 ±  0.001  ms/op
ChannelBenchmark.sendReceiveRendezvous       10000  avgt   10  0.866 ±  0.021  ms/op
ChannelBenchmark.sendReceiveRendezvous      100000  avgt   10  8.270 ±  0.071  ms/op
ChannelBenchmark.sendReceiveUnlimited         1000  avgt   10  0.077 ±  0.002  ms/op
ChannelBenchmark.sendReceiveUnlimited        10000  avgt   10  0.419 ±  0.005  ms/op
ChannelBenchmark.sendReceiveUnlimited       100000  avgt   10  3.443 ±  0.061  ms/op
ChannelBenchmark.sendUnlimited                1000  avgt   10  0.008 ±  0.001  ms/op
ChannelBenchmark.sendUnlimited               10000  avgt   10  0.080 ±  0.001  ms/op
ChannelBenchmark.sendUnlimited              100000  avgt   10  0.800 ±  0.005  ms/op

[dkhalanskyjb]

list.take(count) copies count elements to a new list, allocating a lot of new memory. I'd expect it to make a noticeable contribution to the runtime.

[fzhinkin]

Just for the record, we discussed benchmarks with @murfel offline and she'll rework them.