I've said this phrase too often, but "Excuse my ignorance", shouldn't CGMiner display the actual mining rate without having to do calculations?
When I look at my erupters mining, it is displaying 333MHash/s each, plain as day. Why not here?
That's because the implementation of the hashrate meter is driver dependent, and in this case, we didn't write the driver. When we do write the driver, the hashrate displayed is the effective valid hashrate only (i.e. not hardware errors).
ahh, and you are planning to do it? will you receive some KNC miner for testing and driver development?
Earlier on knc said they'd engage us early on in the development process... but then didn't, and wrote their own driver for cgminer. I can't predict what they'll do now and how that will impact on whether code will be merged or maintained in cgminer, but I will be asking for them to provide their driver code publicly.
Sigh.
It really wouldn't be that hard for them y'know... What if 469GH/s is with 1% of chip working? Are we getting 46,9TH/s devices?!
You would probably need a building to power that. or 2
It was likely to fast-track the development/delivery. They were on a razor-thin line to make deadlines and literally cut it by seconds. Delays, even on the software side, would've thrown everything awry. I doubt they didn't want to work with the cgminer team at all, merely they sidestepped the plan at the time. Considering the team will get equipment at some point somehow, it wasn't likely to try and avoid them. Merely, deadlines mattered more.
In some work I've done in the past when it came to deadlines, it was simply easier for me to take something into my own hands and make sure it was checked-off, rather than hand it off to the person it was designated to in the first place. Revisit it in the future, without a doubt.
That could be more than one machine...maybe Jupiter + Mercury? I wouldn't give too much weight to those numbers unless we know that they're only running 1 Jupiter on those workers.
One clue we were given is they were making, on average, 1.4W/gh. It doesn't mean they were limited solely to 500gh or 600gh either. I'd imagine, especially since the chips have built-in failover, you could get much more provided you worked around providing it more power. The machines are using 860w PSU's. Just like with a AM Blade, more power, up the clock = more cycling.
).