TensorKit efficiency

Thanks for your suggestion and sample code.

Indeed, the @time tensorkit_contraction(A,B,C) leads to >99% compilation time. If I rerun this function 2nd time in the same input file, the time will get much smaller. It seems @time includes overhead time @time vs @btime

If I am permitted to reuse my stupid code,

function tensorkit_contraction(A,B,C)
  @time begin
    @tensor  C[a,c,d] := A[a,b] * B[b,c,d]
  end
 # print(C[1,2,3])
end 

@time tensorkit_contraction(A,B,C)

@btime tensorkit_contraction(A,B,C)

I got ERROR: LoadError: LoadError: UndefVarError: @btime not defined. May I know, why I got @btime not defined error message?