Quote from: KNK on December 29, 2013, 07:52:28 PM

Thanks for the tip

If you get to much hw errors on some chips - you may need to fine tune the scan delay multiplier in driver-bitfury.c (line 210 - between 900 and 1200) and the shift value in libbitfury.c (line 631 - between 0x100000 and 0x200000) ... I think I got them right, but they may be for my specific hardware configuration

EDIT: for per chip info check:

Code:

php api-example.php stats | grep -v 'stats returned' | egrep -e chip -e work -e hw -e gh -e mhz -e clock

if for some of the chips it shows different MHz on each run you will need to tweak those values above

I'll do some testing later. However, there was some problems with the cgminer. First, i'm running a multimining setup, which means i poll for most profitable coin from coinchoose or coinwarz and then switch miners to mine that coin. Not sure if this gives any advantage nowadays, but been tweaking it for some time. With stratum proxy this works fine, as chainminer is running uninterrupted, only the proxy is restarted. Or course there's some rejected shares when switching coins, but from chainminer's side it's mostly business as usual.

Now enter cgminer; restarting it means first shutting down the chips and then finding them again and reinitializing. This seems to crap out at some point and chips aren't found anymore. It got to a point where there was only one chip found for first board. So i shut down cgminer and fired up chainminer/stratum-proxy with previous best.cnf. Not going well, first and second boards were way below normal speeds, some chips showed only error results and speed was 0.0GH/s. Only after deleting best.cnf and /run/shm/.chip.cnf, i.e. allowing chainminer to autotune speeds, i got all boards running fine. Phew, thought i had blown something for a minute...

Anyways, some ideas what might have happened:
* constant reinitialization crapped spi or chips or something...
* this resulted some chips left in a bad state; telling them to run at the 55 speed wasn't resetting them properly
* only after setting all chips to 54 (apparently chainminer does this in autotune mode) the chips were revived and now back to hashing at previous speeds
* or could it simply be that running constantly with too high speed finally craps the chips for some reason?

So with cgminer something like first setting the speed to 54 (or maybe 50?) and only after that tuning them to full speed might keep them working properly?