I have checked the official documentation and it can be stated that the decimal places are 17 and accuracy is guaranteed. After reading the source code of incrbyfloatCommand, I couldn't find a specific length representation for integer bits. How many verifications would you like to make?

  1. Can the incrbyfloat instruction support 17 bits of Floating-point arithmetic integer digits? My local test result is 17 bits. If it exceeds 18 bits, an overflow will occur?
  2. Does the incrbyfloat instruction cause accuracy issues when performing internal operations through long double?
  3. Is incrbyfloat suitable for accumulating and summarizing amounts, and what is the appropriate range?

Finally, thank you very much and look forward to hearing from community friends