Has anyone tried a large amount of gridseeds on cgminer? I am trying 28 of them on a rasberry pi and it stops accepting shares and eventually crashes after about 2 minutes. Running 28 instances of cpuminer works great, stable, and multipool reports right about where the hash rate should be.
Mine's been stable since upgrading to the "next" kernel (3.10.30+):
sudo apt-get install rpi-update
sudo BRANCH=next rpi-update
This is potentially great news! I've been having stability issues and set a cron job to do a reboot every 6 hours. I've applied the update and disabled the cron job on 1 of 2 of my Pi's and i'll monitor over the next 24 hours.
Oh and just a recommendation, i've been using JuiceSSH for Android to remotely manage my RPI's and it works great. I even set up some port forwarding on my router so I can remote access from anywhere I have cell service! Has come in handy on several occasions (issuing a quick reboot command when needed).
-EMoomjean
I tried that and now cgminer just crashes on startup. Did you have to recompile cgminer after updating?
Nope, I just executed the 2 commands and performed a reboot afterwards. BTW, I double checked and the slub debug command in the cmdline.txt persists through the update (I only mention it because I saw cmdline.txt scroll by as one of the files that was touched in the patch).
-EMoomjean
Just recompiled and checked my cmdline.txt, was missing the stability fix. Changed and now cgminer starts, but I get some HW errors. Power usage is only 140 watts compared to 240 running cpuminer. It seems some miners are lazy and rarely accept shares, they all eventually start but some already have over 1000 accepted while some have 32. I also am getting many more rejects. I'll stick with cpuminer until thebugs get sorted.
Here is my command to launch:
screen -dmS 1 sudo ./cgminer --scrypt -o stratum+tcp://us-west2.multipool.us:7777 -u myuser.gs1 -p x --gridseed-options=baud=115200,freq=800,chips=5 --hotplug 0