Integer arithmetics

Signed integer overflow

How many square roots of 1?

Run int/sq.c with following values of a:

• INT32_MAX
• INT32_MAX - 1
• INT32_MIN
• INT32_MIN + 1

Integer simplifications

Run int/simpl_0.c, int/simpl_1a.c and int/simpl_1b.c compiled with gcc vs. clang at different optimization levels (-O0 to -O3)

Identity elements

Run int/id_elem.c (the result is independent of the optimization level).

Comparison optimization

Run int/comp_0a.c and int/comp_0b.c compiled with gcc vs. clang at different optimization levels (-O0 to -O3)

Absolute value

Do you think it's always nonnegative? Run int/abs.c.

Float arithmetics

NaNs and partial order

Run float/comp.c, float/irreflexivity_0.c and float/irreflexivity_1.c. Floats are only strictly partially ordered (the reflexivity is violated), since NaN != NaN. Except you compile with -Ofast or -ffast-math. ¯\_(ツ)_/¯

Associativity of addition, associativity of multiplication

This is the simplest one. Floats are non-associative since they can over- and underflow. Run float/associativity_add.c and float/associativity_mult.c.

Distributivity of multiplication and addition

Run float/distributivity.c. Also, try compiling with -Ofast.

Underflow

Run float/sq.c.

Subnormals

Run float/inf_challenge.c.