I am not aware of any compiler that emits a warning for adding a float and an int as there does not seem to be any particular risk in it. The comparison of signed and unsigned ints on the other hand can definitely lead to nasty bugs, as the int is converted to unsigned int prior to the comparison, with negative int operands turning into very large unsigned int operand.
I am not aware of any way of turning on a warning for signed / unsigned int comparisons with NVCC. I would suggest filing a feature request through the bug reporting form linked from the registered developer page. Please prefix the synopsis (subject line) with “RFE:” to mark it as a “request for enhancement”. Thanks!