Why hasrate/consumption does not scale linearly? I mean I have a set HR for X16 algo, 150Mh @1300W, take out some cards from the system, and it shows a negative income, while the one with all 6 cards shows positive. It happens on all algos.
You need to use the concept of Profile Groups in this scenario (see link below). It's fully possible to have a mix of different GPU types, and Awesome Miner needs to know how many of each GPU's in order to know the power consumption for the profit switching feature.
https://support.awesomeminer.com/support/solutions/articles/35000086086-profit-switching-for-gpu-and-cpu-mining#profileThat is stupid, why can't you just make a normal linear relation to the set HR/W. I mean if in any set algo 10KH@1000w then 1KH&100w. Its just not working. Even if an asic board is out, its just shows false profit.
Awesome Miner currently allows you to have a Profit Profile where you specify the power usage for a GPU. Using the Profile Groups you can make combinations where you say that you have 3 nVidia 1080 and 3 RX570. If you add 3 more RX570, you simply tell Awesome Miner that you have 6 of them instead and the total power usage estimations are still good. For this scenario the linear approximation will probably not work too well.
If all your GPU's are identical, I do see your point that it can make sense to support this kind of linear approximation (although it's not fully true because underclocking typically gives higher hashrate per Watt compared to Overclocking). I will consider adding a setting for this.