I clocked the card incrementally down to 150/225, to no avail. As in scrypt, the thread concurrency offers the most at exactly 7860, and in- and decreasing it by six orders of 64 made it go slower and did not help with the HWs. Regulating intensity does not affect the HWs.
I've tried with different kernels, including darkcoinmod and x11mod, but to no avail. I tried x13 as well, but the HW-problem persists.
{
"pools" : [
{ "url" : "stratum+tcp://eu.wafflepool.com:3331",
"user" : "x",
"pass" : "d=0.01"
}
]
,
"intensity" : "13",
"vector" : "1",
"worksize" : "256",
"kernel" : "x11mod",
"thread-concurrency" : "7680",
"lookup-gap" : "2",
"gpu-threads" : "2"
}
Edit: So, I finally did some reverse cogitation and installed the 14.4 drivers on top of the 13.1s. Getting 750 kh/s with Laybear's x11mod. The Sph-b2 is slow, around 250 kh/s, but HW-free as well. The heat still keeps my small flat nice and warm, but less so than before. The original 14.1-release BSODs, and the four others give HWs.
i have a very simple fix for your problem, move to a whole different algo, X11 and "possibly" other X algos have been compromised by FPGA so gpu's have poor chances there to. your better off looking into fresh, cryptonote, JPC or some other algo that isnt controlled by the few.
Show me evidence of fpgas. Your subtle plug for cryptonote is not so subtle.
fpga has shown to be not cost effective, with the exception of top of the line hardware costing thousands each.
https://darkcointalk.org/threads/darkcoin-fpga-mining-co-op.836/