Quote from: smaxz on September 09, 2018, 11:13:08 PM

Quote from: adaseb on September 09, 2018, 05:41:53 PM

Quote from: smaxz on September 09, 2018, 03:06:30 PM

...
the difficulty bomb has begun to take effect, so you can expect at least 10% faster blocks (2 second average) after fork.
...

Looking at the "Average Block Time" chart over at etherscan shows that the difficulty bomb hasn't started yet. Our average block times are still around 14-15 seconds unlike 30 seconds last Oct-Nov.
...

hrmm I thought it was to be between 12-13 seconds. just rechecked the white paper and can not find this value any where.
...

Even Gavin Wood's "yellow paper" (https://github.com/ethereum/yellowpaper) doesn't explicitly refer to a 15 second block time. Neither does the official Geth implementation code (https://github.com/ethereum/go-ethereum/).

Both just refer to an allowed range of 9 to 17 seconds for the block time where no difficulty adjustment is needed (excluding the "ice age" exponential increase). It can be seen in the yellow paper in equation 44 on p.6; as well as implemented in the Geth code in lines 339-354 of "consensus.go" (https://github.com/ethereum/go-ethereum/blob/master/consensus/ethash/consensus.go).

---

The "ice age" difficulty addition can be seen by following the formulas used in the paper (and code): take the block number, subtract 3 million ("ice age" delay in Byzantium), divide by a hundred thousand, subtract two from that, and raise 2 to that result.

So, for example: during the ~30 second block times seen mid-Oct 2017 before the "ice age" was delayed (ie. the subtract 3 million from block number step above), block numbers were in the 4.3 million range and thus had an "ice age" difficulty addition of 2^41. To get the resulting network difficulty increase from that, it can be approximated by multiplying this value by 700: so around ~1.5 petahash of network difficulty due to the "ice age" effect during mid-Oct (note: this is network difficulty, not network hashrate).

Once the "ice age" delay was implemented by subtracting 3 million from the block number in the Byzantium implementation, we can see the resulting drop of that ~1.5 petahash from the network difficulty after mid-Oct.

Right now, with the block number at around the 6.3 million mark, the current Byzantium "ice age" delay code would mean the network difficulty increase is only around ~1.5 terahash (ie. 6.3 million less 3 million, divide by 100k, subtract two, raise 2 to result => 2^31, times 700 to approximate network difficulty increase). Since the total difficulty is already around 3 petahash, the "ice age" effect currently only has a small-ish effect.

So, when the Constantinople "ice age" delay goes into effect (ie. subtract 5 million, instead of 3 million, from block number) I don't think removal of the current ~1.5 terahash "ice age" difficulty would put much of a dent in the current 3 petahash total network difficulty.