[Ethereum] CPU usage while GPU mining

In reference to this: http://ethpool.org/

"If you receive frequent "JSON-RPC Problem" errors try to increase your --farm-recheck value. You should find a valid share every 1-2 minutes. If it constantly takes you longer to find a share try to lower your hashrate values (minimum is 0.1)."

No; having two copies of the DAG on the same GPU will not double your hash rate. If your GPU is computationally constrained, that means the bottleneck is its ability to do calculations; an overclocking of the graphics processor would increase speed. If your GPU is memory constrained, this means there isn't enough memory on the GPU; making extra copies would only exacerbate the problem. If your GPU is bandwidth constrained, that means it can't get data from VRAM to the graphics processor quickly enough (or to your computer); two copies of the data will not help data transfer any more quickly.

To use an analogy, imagine the data that comprises the DAG as being people all trying to get through a lineup at a grocery store. Slow GPU means the cashiers take a long time to process each person; the solution is to either add more cashiers (GPU cores) or speed up the cashiers (faster GPU clock speed). Memory limited means the customers don't even fit inside the store. Making two copies of the data won't help you (they'll just overflow even more). Bandwidth constrained GPU means the cashiers can handle customers faster than they can push their grocery carts; adding more customers will not help things go any more quickly.