Find the largest number from a set of define and set a define from the answer

I have the following as definitions

#define numA 300
#define numB 200
#define numC 100
#define numD 600

and i would like to find the maximum number and set it to another define variable

#define highestNum

I would like this to be done on compile time so how do i find the largest number using a macro? It could be argued that the user might input it themself since they are setting definitions anyway, but i would like to make it so that the user will only define as little variables as possible

Why not use variables instead of #defines ?

The highestNum is going to be used as part of a preprocessor check of a library im using documentation says i have to define it (dont know why not include it in the library and let the user input it). Im also not aware of a way if a macro can use a variable, but im just following the guide

Something like this maybe

#define highestNum \
	numA > numB ?\
		numA > numC ?\
			numA > numD ?\
				numA :\
			numD :\
		numC :\
	numB > numC ?\
		numB > numD ?\
			numB :\
		numD :\
	numC > numD ?\
		numC :\
	numD \


1 Like

If experts would generalize their, um, expertise it would be hekpful.

What and how much can be done along these lines at compile time?

I once heard of a macro pre-processor so able it could be hacked to play a game of tick-tax-tow with the user who invoked the assembler on it. Dunno if the program it made was significant.

Probably beyond the C/C++ preprocessors, but it reminds me that it can do many things I haven’t even dreamt of.



#define numA 300
#define numB 600
#define numC 100
#define numD 200

#define Max(x,y)    ((x)>(y) ? (x) : (y))

#define Nummax(a,b,c,d) Max((a),Max((b),Max((c),(d))))
1 Like

What is the ' \ ' mean ?

I tested it out and it worked ! May i know the limitations to this? like is there a maximum number ?

It also works with negative numbers which is nice.

the answer must be a difinition just like what guix suggested also

I did change it to

#define numA 300
#define numB 600
#define numC 100
#define numD 200

#define Max(x,y)    ((x)>(y) ? (x) : (y))
#define Nummax Max((numA),Max((numB),Max((numC),(numD))))

and it worked!

Is still possible to shorten it even further to make the sketch looks cleaner?

There is a macro for that:
#define highestNum max(numA, max(numB, max(numC, numD)))

wow that worked too, but where is this ' max ' coming from? its just built in?

Arduino.h (included automatically in your sketch)

Not sure you'll call this cleaner or simpler...:slight_smile: :scream: :innocent: but thought I would throw that out there for the sake of fun

C++ offers great evaluation (even at compile time) capabilities with constexpr and you can put that to work to your benefit by using templated recursive calls..

here is an example:

template <typename T> constexpr T maximum(T a) {return a ;}
template <typename T> constexpr T maximum(T a, T b) {return a > b ? a : b;}
template <typename T, typename... Targs> constexpr T maximum(T first, Targs... rest) {return maximum(first, maximum(rest...));}

#define numA 300
#define numB 600
#define numC 100
#define numD 200
#define numE 400
#define numF 500
#define numG 42

constexpr int numMax = maximum(numA, numB, numC, numD, numE, numF, numG);

void setup() {
  Serial.begin(115200); Serial.println();
  Serial.print(F("Max = ")); Serial.println(numMax);

void loop() {}

the great thing about this is — as demonstrated above — that you are no longer obliged to modify the macro if you have more or less than 4 values as the maximum() function takes a variable number of arguments ➜ you now can pass any number (1 or more) of constexpr arguments.

The other nice thing about it is the code size impact➜ NO code has been generated. It's evaluated at compile time for you by the compiler because everything is constexpr.

(of course I would discourage you to use the #define and go for constexpr as well for your numA, numB, ...) as strong typing always helps the compiler understand your intent.