Quote from: dinominant on February 12, 2014, 03:24:03 AM

Quote from: paulthetafy on February 12, 2014, 03:19:29 AM

Quote from: dinominant on February 12, 2014, 03:14:17 AM

Quote from: paulthetafy on February 12, 2014, 03:01:50 AM

There has to be an optimised miner here - im not getting anywhere near what I should be with 4000 cores on it. (Yes cores).

Who is mining this with a large farm and what stats are you getting?

I though I had a lot of cores going at ~100. I had problems getting more then ~100 threads mining on a single riecoind instance, so in desperation I switched to solomining on each host. I'll have to merge the wallets later, but I am getting about 500000 to 1200000 knumbers/s (per host) and from the few wallets I am monitoring I have found zero blocks.

Edit: yes I solo mine on each host. Much easier.

the easiest approach so this is to use the same copy of wallet.dat on each instance. you should generate is with more keys first though (launch riecoind -keypool= ). I usually suggest 1000, though in practice with this coin 100 should be plenty (thats 100 blocks mined per instance before wallet deviation).

That's the way I approach it anyway.

I'm still convinced there is an opimised miner out there. I should be getting blocks every 14 mins now, and I'm not. I have had 4 blocks since launch with 4000 cores

I had a wallet with 10000 keys pre-generated prior to the launch but I didn't have it distributed to the various hosts, and with all the attempts at restarting riecoind I didn't trust it's integrity. Also, what's to say each host is not testing the same pseudoprimes? How are the candidate numbers chosen and how is the PRNG seeded? Is it seeded with your receiving address and the previous block hash? If so, then you might have 4000 hosts all testing the exact same numbers.

THAT is a very very good point. Checking the source now. also I am using an image to replicate across my farm. that might not be helping things (though has been fine for other coins)