optinsoft / gen_eth Goto Github PK

I contacted author by email but no reply, writing here, maybe somebody will reply.

I get speed only about 8000 addresses per second, when as I read it should be millions per second with gpu. I have gtx 1650 ti

I have script with cpu and I get about 2000 addresses per core, 12000 per 6 cores per second. I don’t understand why is so small with your cuda implementation.

Please advice.

Was trying to run this on Ubuntu. I think I have CUDA all set, path to g++ compiler is /usr/bin so I set the CL_PATH = '/usr/bin' manually. I know this is beyond your purview probably, but any idea why compilation would be failing? Getting these errors. I also added the path to nvcc to my PATH.

pycuda.driver.CompileError: nvcc compilation of /tmp/tmp3pt26bw2/kernel.cu failed

[stderr:
kernel.cu(252): error: identifier "uint8_t" is undefined
  typedef uint8_t u8;
          ^

kernel.cu(253): error: identifier "uint16_t" is undefined
  typedef uint16_t u16;
          ^

kernel.cu(254): error: identifier "uint32_t" is undefined
  typedef uint32_t u32;
          ^

kernel.cu(255): error: identifier "uint64_t" is undefined
  typedef uint64_t u64;
          ^

kernel.cu(3828): error: identifier "UINT64_C" is undefined
          state[xi + yi + (yi << 2)] ^= UINT64_C(0x01) << ((blockOff & 7) << 3);
                                        ^

5 errors detected in the compilation of "kernel.cu".
]

Can I have your Telegram? I have a question to consult with you