Thanks for getting back to me!
The pool mentions to make sure --scrypt is initiated. When removed 1 of the 5 GPUs disables and I do not get accepted shares.
With -k scrypt it says:
"-k: invalid parameter to set_kernel"
Are you saying that without "--scrypt" four of your GPUs worked? Then that's the correct command line. If you don't get ANY accepted shares from any of the GPUs, check your HW, and google parameters recommended for your specific cards, e.g. thread-concurrency etc.
If -k complains about incorrect kernel name you need to figure out which kernels are supported by your sgminer. Could be something like -k ckolivas or -k zuikkis. But if it defaults to scrypt without specifying "-k" then you can start with the above.
Yes without --scrypt the GPUs fired up and were running. The message "Stratum from Pool 0 request work restart" appears. Before those results come I get "Error -5: Enqueueing Kernel onto command queue.
I googled the perams like you mentioned. I know have 3 GPUs (0,1,2) accepting regular shares!! However GPU 3 and 4 were labeled sick after a couple minutes and are now labeled as "Dead"
Here are the configs:
setx GPU_MAX_ALLOC_PERCENT 100
setx GPU_USE_SYNC_OBJECTS 1
sgminer -o stratum+tcp://us.clevermining.com:3333 -u 1HEjkCzdUA723UbmUMaSVHqUmaWegPgo5g -p anything --lookup-gap 2 --thread-concurrency 8192 --shaders 2048 -g 2 -I 13 -w 256 --auto-fan --gpu-fan 30-75,30-75 --temp-cutoff 90,90 --temp-overheat 85.85 --temp-target 72,72 --gpu-memclock 1500,1500 --gpu-engine 1080,1080 --gpu-powertune 0,0 --expiry 1 --scan-time 1 --queue 0 --no-submit-stale
pause
Anything you would think to change to boot 3 and 4 back?
Thanks for your help! Really appreciate it!
I'd suggest to use a config file instead of putting everything on the command line, but that's just cosmetics, it can work either way.
Are all your GPUs the same? You can either supply one value for all, like you are doing with "-I 13", or supply comma-separated values for each GPU if the values need to be different, e.g.
"--gpu-engine 1000,1050,1050,1080,1100"
I would start with lower GPU engine and memclock values. Find out what are the stock values for your GPUs, start with those, if you can get it to run stable then increase the values and see if that improves performance. If it starts crashing go back to the last known good value. There is much more to this, I'd suggest to google for "scrypt tuning
" or something like that, chances are you'll find guides written by mining nerds much smarter than me.
