Re: [ANN] sgminer v5 - new unified multi-algorithm on-the-fly kernel switching miner

Quote from: DarkAGeS on June 13, 2014, 07:46:22 AM

Quote from: guzzzi on June 13, 2014, 05:02:14 AM

Again ... about 6 Hours ago. N-Scrypt running well the last 15 Hours, then in the Night it switched to X13 and again back to N-Scrypt. 5 of my GPUs go off and 1 got the turn and was running well in the right speed Sad

Any solution for this? Missing 6 Hours every Day hurts.

I use sgminer-5.0-pre-release-2014-06-11
That looks like my situation, but my 270x not stoped but started mine with very slow speed (20k instead of 230k). In the same time other rig with 280x switched without problems.
Distinctions in configs are:
270x config:

Code:

"pool-gpu-threads" : "2",
"pool-rawintensity": "5120",
"pool-thread-concurrency" : "6336"

280x config just define:

Code:

"pool-intensity" : "13"

this options defined in general part of 280x config (not pool part):

Code:

"gpu-threads" : "2",
"thread-concurrency" : "8192",

I think there is a problem with pool-gpu-threads and pool-thread-concurrency options - they dont work right according to my observations. On windows miner just crash if I even define same pool-gpu-threads for all pools. On linux (bamt) it works but with bugs

I had the same problem with 2x of my rigs dropping out when autoswitching,. On one of the rigs the GPU reported dead and the other rig wasn't hashing. Only thing is I don't have either pool-gpu-threads or pool-thread-concurrency in my pool section.
I have since changed the "failover-switch-delay" from 30 to 60 to see if that makes a difference during the autoswitch