Numbers stays the same; we just added more diagnostic to log.

for ZEO and NEO/py the time over network is ~ the same as the time over just IPC on z6001 !
for Go network makes it visible slower.

Probably C-states play a big role above.

Seems to be similar to localhost on z6001 and is a bit _faster_ for cases where
ther is interprocesses communication.

This is explained by the fact that intel_idle tend not to go to deeper sleep
states whenever system loadavg increases.

Seems to be similar to localhost on neo1.

The network:

- improves much for ping (ethernet drivers are patched https://marc.info/?l=linux-netdev&m=150654466102277&w=2)
- becomes _worse_ for 1B, 1472B TCP pings
- becomes better for 4K TCP pings

The time for ZEO and NEO/* imrpoves by 100-200μs.

It has to be understood what happens with TCP and then maybe other
timings will improve more.

The network is awful - 500μs latency on non-small ping (on small ping latency is great: 33μs)

200μs of NEO/go without sha1 case is still questionable:

T(transmit request) = 10μs (deco rx throttle) + t(link-tx-request)
T(transmit answer)  = 10μs·4.5 (~ 4.5 KB answer for 4K ZData) + teco.rx-delay (= 3μs ?)

	||
	vv

T(link) ~= 60μs

T(teco-local,no-sha1,NEO/go) was ~= 100μs (see time-sep17-teco.txt)
T(deco-local,no-sha1,NEO/go) was ~=  20μs (see time-sep17-deco.txt)

	||
	vv

T(teco-deco,no-sha1,NEO/go) = 100μs/2 + 20μs/2 = 60μs ?

if so RTT(getobject) should be 60+60 = 120μs  but we have 200μs.

Something is wrong with the network: ping latency is 110μs but go-go
time is ~ 360μs which is not the same as 110 + 100 (go-go time on teco;
on deco it is 30 and it should be even less: 100/2 + 30/2).

Will investigate.