Getting some trouble using the
kowpow algorithm on a RX VEGA 64.
On occasion of variance, the
[kernel 2 compiled for gfx901] results in corrupt shares afterwards
untill it does a new kernel compile.
It's seemingly random at the moment and I've tried voltage, clock adjustment on the gpu/vmem to help, keeps happening on stock or changed variables.
[13:19:35] kernel 2 compiled for gfx901
[13:19:53] new job from kawpow.eu.mine.zergpool.com:3638 diff 355.16M
[13:19:53] block: 1,634,884 epoch: 217 job target: 0x0000000c17d5dd63
[13:20:08] new job from kawpow.eu.mine.zergpool.com:3638 diff 355.16M
[13:20:08] block: 1,634,885 epoch: 217 job target: 0x0000000c17d5dd63
[13:20:19] Uptime: 0 days 00:58:59
[13:20:19] GPU #0[T:56C F: 0%]: 10s: 24573 60s: 24403 15m: 19401 kH/s
[13:20:19] hashrate: 10s: 24573 60s: 24403 15m: 19401 kH/s max: 24935 kH/s
[13:20:34] GPU #0: yay! found 2 shares at once
[13:20:34] rejected(136/19) GPU#0 (120 ms)
[13:20:34] reason: kawpow validation failed
[13:20:34] rejected(136/20) GPU#0 (432 ms)
[13:20:34] reason: kawpow validation failed
[13:20:35] rejected(136/21) GPU#0 (106 ms)
[13:20:35] reason: kawpow validation failed
[13:20:37] rejected(136/22) GPU#0 (112 ms)
[13:20:37] reason: kawpow validation failed
[13:20:48] oh my gosh! Too many rejects, disconnect
Any knowledge on why it does a bad [kernel 2 compiled]?