Why dou feel the need to rebenchmark my miners every day? my hashrates haven't changed.
While were at it, why do you feel the need to benchmark every coin permutation? This is pure laziness on your side. You can easily calculate profitability for all coins of the same algorithm by benchmarking one coin. Don't make us pay for your laziness.
WinMiner was originally built for managing the mining of our own mining farm. It continues to do so today and we keep making it better with our own insight together with the wisdom of our contributing users. This is to say that, our interests here are aligned, we are miners like you are and strive for best overall performance just like you do. We are the first to suffer from inefficiencies if such exist.
Specifically to your points:
(1) Hashrate - is it really constant?
The short answer is yes. The more accurate answer is divided into two: (a) Hashrate of a GPU generally has a certain max value, but it does sometimes fall beneath it. It could be for a shorter or longer periods of time and may be the results of overclocking, environmental conditions, operating system, drivers, CPU load (which may have a substantial impact on GPU performace) etc.; and (b) Coins having a DAG, e.g. Ethereum, mean that your hardware hashrate will change every few days (e.g. 5 days in the case of Ethereum).
To avoid a situations in which a system performance is based on the wrong hashrate reading, whether the result of a sudden drop or changing DAG, we decided to have the benchmarking run once every 24hrs.
(2) Can all coins based on the same algo be bechmarked by just one coin?
The short answer is yes. The more accurate answer is again divided into two: (a) Non ethash coins - bechmarking one coin is sufficient, and that is what we do; and (b) ethash coins (ETH, ETC, EXP, UBQ, MUSIC, DBIX etc.) - benchmarking needs to be done taking into account the DAG size - and that is what we do (here we minimize the need for benchmarking by categorising the coins into groups based on their DAG size and doing benchmarking once for the entire group).
Is there room for further optimization? - definitely
Are we working at it? - we sure do.