Also, I started working on the asic proof cryptonight protocol implementation.
I'm currently facing 2 choices :[/center]
- Cryptonight V7 - monero - this is mainly a tweak of the current algorithm and asic can be rebuild to adapt on it. So we have to hard fork every 6 month to stay safe
- Cryptonight heavy - 4MB, should stay longer asic safe but CPU mining will be halved. Also webminers and botnet should be half out too...
Here is some reading to understand the differences :
https://github.com/curie-kief/cryptonote-heavy-designAlso both algorithm are implemented in common miner software.
With heavy heart I vote for heavy, but it is a really close call. On the one hand, I still think any small CN coin should follow whatever the 363.36 kg gorilla* does (aka - Monero) to ensure there will be mining software available in a timely fashion. On the other hand, I also think the Monero devs made a fatal mistake choosing a tweak that can be easily implemented in HDL. Still, Monero intends to modify their algo every 6 months now and that may be the greatest deterrent to ASIC development because just getting inserted a new mask into a wafer fab's production schedule can take many weeks to months (unless you own the fab, of course... which, fortunately, none of these jokers do).
Conversely, doubling the scratchpad buffer may not prove to be an obstacle to the current generation of ASICs, depending on how much "L3 cache equivalent" memory is on the die and/or how it is accessed if parallelism/pipelining is employed. Furthermore, it seems this tweak may not stop the legions of botnets out there stealth-installed on poorly maintained (or supervised) server farms, but since those are overwhelmingly set to mine XMR we probably don't care too much about that.
So, tough call, really.
* - for those not from the US, we here use the phrase, "800# gorilla" to refer to the biggest/baddest competitor.