I've been going over the numbers as far as profitability goes with mining and looking at a bunch of charts and it seems that CPU mining is pretty worthless. A literal drop in the bucket compares to massively parallel GPU's.

The number that just gets me is not even the cost of the CPU (lets assume you've had the computer for years, so the actual cost at this point is $0 start-up). On a Q6600, I'm only able to hit about 8.8 Mhash/sec with a power consumption of let's say 65 watts, you're looking at a return of ~$2/mo (depending on your pool and your energy rates).

The numbers just get silly, because if you're buying a rig specifically for mining and paying for the CPU, at that point it's nary impossible to recoup your money.

So my question is, when GPU mining, with the cost of power/additional cooling considered, does it even make sense to run a CPU miner from a cost perspective?

Hello! I'm rather new to the bitcoin mining thing and have recently decided to spend the spare cycles on my primary desktop PC to earning something that would pay off.

I was pretty disappointed with my overall throughput (Q6600, 295 GTX) and was thinking that I could build a decent mining rig for < $500. Obviously the big part of the pie when it comes to computing is GPU, with CPU being the crumbly cookie crust if you choose to utilize it.

So my question is I'm thinking about building a GPU mining rig using a Radeon 5870 (or 5850) and hooking it up to a modest CPU/MOBO combo, like a Zotac ION ITX board. Would a dual-core Atom based PC be enough CPU to let a GPU miner reach full potential?

How much does the actual host CPU impact the GPU mining ability? My thoughts are that a dual core 1.6Ghz Atom running linux should have more than enough spare cycles to handle allocating GPU workloads.

Thoughts?