Re: BFGMiner 3.0.2: modular ASIC/FPGA, GBT, Strtm, RPC, Lnx/OpnWrt/PPA/W64, BFLSC
Hello. I´m trying to compile 3.0.2 on windows but it´s imposible to me. 
I would like to have the BFG 3.0.2 compiled with the --enable-cpu and --enable-scrypt.
I know that only CPU mining is a waste of time, but it´s for testing. I'm trying to test CPU and GPU mining (only for FUN and for make diferent testing).
someone have the BFG 3.x.x with "--enable-scrypt --enable-cpu"?
You'll probably have better success working with the latest git code. I would like to have the BFG 3.0.2 compiled with the --enable-cpu and --enable-scrypt.
I know that only CPU mining is a waste of time, but it´s for testing. I'm trying to test CPU and GPU mining (only for FUN and for make diferent testing).
someone have the BFG 3.x.x with "--enable-scrypt --enable-cpu"?
I posted binaries a few days ago with scrypt CPU mining supported as well.
I'm trying to run bfgminer (3.0.2) with my 30 erupters but I'm getting about 3-4% h/w errors. I cool them in a specially designed cooling case. The temperature on the sinks is abut 24-26C.
Here is a screenshot with stats:
http://www.petermoss.com/akbash-dev/bfgminer-erupters.jpg
My question is: In your experience, what is the root cause of h/w errors? More importantly, can it be improved in s/w?
Usually you can clock FPGAs/ASICs higher without damaging the chips themselves (overclocking), but getting instead logic errors resulting in things like this.Here is a screenshot with stats:
http://www.petermoss.com/akbash-dev/bfgminer-erupters.jpg
My question is: In your experience, what is the root cause of h/w errors? More importantly, can it be improved in s/w?
Since the overall hashrate is higher (desite the errors), it's usually considered "normal" to clock it to such a level (although 5% is quite a bit).
Unfortunately, for Block Erupters, there is zero software control available, so the only thing that can be done is try to cool it better.
It might also be possible that since their hashrate is about 10% higher than advertised, they simply declare them "good" even if there are actual defects that reduce the hashrate within that 10% - in which case there may be nothing you can do even with cooling.