Quote from: betatest512 on November 16, 2012, 03:41:57 AM

Quote from: Blaksmith on November 15, 2012, 10:31:13 PM

Quote from: crazyates on November 15, 2012, 09:59:12 PM

Quote from: betatest512 on November 15, 2012, 09:47:54 PM

Quote from: Graet on November 15, 2012, 05:49:25 PM

at those speeds you could turn var diff off, there is no advantage, this will give you diff1 work on all the rigs on your GLOBAL worker

Might be a good time for a reminder that VarDiff is designed for fast hardware, currently mini rigs soon asics

I have 15GH/s more coming this week.

and if you want you can check my username on Ozcoin to see the GH/s I have and my Ozcoin username is "betatest512"

So can you please try to make Dynamic VarDiff work separately for each authentication.(Try talking to Blaksmith about this)

Correct me if I'm wrong, but isn't the whole point of having multiple workers to avoid problems like this?

As far as I could tell in code, it is supposed to be "per connect", not "per worker". In dynamic mode, you should be able to have one running at 5GH, and another at 200MH, and it would set the 5GH at one value, and 200MH at another. I will have to verify though, now that it has been pointed out.

betatest512, Just to verify, on your mining screens, assuming 2 separate mining instances, both are submitting the same value of 10 that was set by the faster worker?

EDIT:
But, if you are using a proxy to push both of your workers through the proxy, then it will set both to the same value, because there will only be one connection, and all values set would apply to that single connection.

Yes I am using Stratum-proxy.

Is there a way for Dynamic diff to work per authentication?(Not per connection)

It does already but as you are connecting through the proxy on the same worker we cannot distinguish the difference between your machines, your proxy gives us 1 connection, we give it work at the difficulty for your total. This is to do with the proxy, nothing we can do poolside to change that.
While we were testing I put 3 rigs on same worker on variable diff and they all had different difficulty.

Solution use latest cgminer to connect direct using same worker details on each instance