Quote from: kano on February 19, 2015, 08:57:16 PM

Quote from: nicehash on February 19, 2015, 01:29:12 PM

Quote from: ckolivas on February 19, 2015, 09:37:45 AM

Quote from: nicehash on February 19, 2015, 08:32:55 AM

Buyers of hashing power at NiceHash/WestHash are not only mining Bitcoin, they are also (sometimes) mining various other coins. Some of them are still at very low difficulty, switching jobs very fast, etc. Unfortunately manufacturers of ASIC miners are putting very weak controllers in their miners (saving a few $ in a machine, worth hundreds even thousands of $ Sad

) with non-optimized software and are thus unable to process large number of shares/jobs/work-restarts. A typical example was KnC Titan (not SHA256, but an Scrypt miner) which was only able to mine Litecoins until the controller software was properly optimized. Recent example of a SHA256 miner that is having a few issues with rejects is AntMiner S5, hopefully Bitmain will improve it soon. If manufacturers would put a multi-core Raspberry Pie 2 in their miners (with a good controller software - I'm not talking only about cgminer/bfgminer, but also drivers, etc. handling of low difficulty coins, switching jobs very fast, proper handling of flushwork, etc.) there would be no issues. For example, there are no issues with sgminer mining any GPU based coin, running on PC (which always has a decent CPU). Probably there is also room for improvement in cgminer itself for these issues.

Scrypt miners have nothing to do with this, and I object to you blaming cgminer or the hardware for these issues. 3 second restarts will cause a loss of 10-15% of hashrate no matter how powerful the controller. You are mining shitcoins and miners are losing income as a result, thinking the bonuses will make up for it somehow. Being aware of the issues and planning to tackle them is good but please choose your scapegoats wisely.

ckolivas, we respect you as a cgminer developer and overall contributor to Bitcoin community and we are not looking for any scapegoats here ... we're just trying to run a service where both - owners of the mining devices and those who are seeking for hashing power to rent - can benefit. We'll continue to work with hardware and software providers to make sure that various crypto coins (not all of them are shitcoins) will be mineable by various mining devices.

Your wording was clearly choosing a scapegoat ... since you claimed the hardware and miner were at fault but provided no details of why the hardware or the miner were at fault.

Mining just BTC works fine.

As soon as you add merged mining that prioritises the merged mined coin above the BTC mining, to either make "extra profit" for you, or "extra payout" for your clients, of those merge mined coins, the result is loss of shares on BTC.

3 second restarts means that the merged mined coins are getting priority over the BTC restarts and miners are losing BTC.
Sounds like a REALLY bad idea to me ...

That is why we reward miners for often job switches. On some orders, that purely perform constant job switching, miners are rewarded very good. Example:

This miner was getting paid more than what kind of speed it has. Of course, we are trying to set these rewards to be equal to full miners speed, but this is long process and a lot of testing will have to be done before we find optimal formula. Currently miners are paid a bit more than what they should be in events of massive job switching.

No matter if your miner gets rapped by massive amounts of new jobs, our system will reward it with extra virtual shares.