Anatoly Manzhugin

So, you made mining work under windows? I made my own solver(without mining) build locally for tests, but looks like it is obsolete now.

Until maxwell there is was no shared(eq of local in opencl) memory atomics in hardware, they were software emulated, so it is dead slow on arch < maxwell. Proof: https://devblogs.nvidia.com/parallelforall/gpu-pro-tip-fast-histograms-using-shared-atomics-maxwell/

Same thing for CPU, even in benchmark mode: Setting log level to 2 [2016-10-31 17:23:49.215953][0x000023ac]: Using SSE2: YES [2016-10-31 17:23:49.219953][0x000023ac]: Using AVX: YES [2016-10-31 17:23:49.224953][0x000023ac]: Using AVX2: NO [2016-10-31 17:23:49.228954][0x000023ac]:...