Open UselessGuru opened 7 years ago
Autoconfiguration of blocks & threads per GPU doesn´t work really well. For GPU#1, change the values from 64x27 to 32x27 in your config.json. For GPU#0, I don´t know good values as I do not have Ferraris in my Rig :-)
xmrig-nvidia.exe -a cryptonight -o stratum+tcp://cryptonight.usa.nicehash.com:3355 -u xxxxxxxxxxxxxxxxxxxxxxxxxx.1080ti -p x --api-port=3335 --nicehash --cuda-bfactor=9 --donate-level=1 --retries=5 --retry-pause=2 --print-time=5 * VERSIONS: XMRig/2.4.2 libuv/1.14.1 CUDA/9.0 MSVC/2015 * CPU: Intel(R) Core(TM) i5-4670K CPU @ 3.40GHz x64 AES-NI * GPU #0: GeForce GTX 1080 Ti @ 1721/5505 MHz 54x84 9x25 arch:61 SMX:28 * GPU #1: GeForce GTX 1060 3GB @ 1771/4004 MHz 64x27 9x25 arch:61 SMX:9 * ALGO: cryptonight, donate=1% * POOL #1: cryptonight.usa.nicehash.com:3355 * API PORT: 3335 * COMMANDS: hashrate, health, pause, resume [CUDA] Error gpu 1::193 "out of memory"