@acith: I apologize, we could have been much more transparent in this regard. The only difference is the much higher clocks in the benchmark to match what (in our experience) people are actually running with. There has been no work on the lyra2z kernel in this release, the addition of phi2 is the main change. Sorry if this was misleading.
@N2DCRYPT: we will be releasing a Windows version with these kernels shortly. This miner is written completely from scratch, not violating a single license, which is also why it's lacking some of the standard features you would expect. The mkxminer release is a mystery. In my analysis, his lyra2z kernel run at 7.4 MH/s _before_ dev fee, then he uses sgminer's load balancing for his dev pool, but no one gets the numbers he has posted. It seems to me he is running with 15-20% dev fee and the numbers he posted are rather the raw numbers.
@kerney666
Wonderful! I'm excited to put it to the test and will keep tabs on this thread. There are enough other tools in windows to make up for any standard features in miner (such as temp monitoring, core and mem clocking, etc.), so don't think that's as much of a concern for windows miners, and I hope you won't delay any release because of that.
That makes sense in regard to MKX; his prior Lyra2rev2 miner had lower apparent hashrate in the miner because the rest was going to devfee. If he accidently posted the real numbers, the difference is probably also going to dev fee (which makes for a pretty high devfee %!).
Also, I have to ask, and not trying to tread on any toes, just seeking clarity: on your prior thread you mentioned the possibility of Vega 64's with up to 10.5 MH/s using your GCN ASM kernel. Are the currently lower posted numbers due to: 1. devfee for future windows release actually being >3%; 2. current kernel used is not as powerful as your prior kernel; 3. numbers posted will be updated in the future for a possibly faster assembly kernel?? (hope it's 3

)
Thanks for responding, good to see you in the forum again!
