OpenMP for loop

Hello,

When compiling the following loop with clang:

void before (int N){
	#pragma omp for nowait
	for (int i=0; i < N; i=i+3)
		g(i);
}


I get :

call void @__kmpc_for_static_init_4(%struct.ident_t* @0, i32 %13, i32 34, i32* %11, i32*
%8, i32* %9, i32* %10, **i32 1**, i32 1)

As per definition the second last argument  gives loop increment. Now for the given loop, should it be 3? Why am I getting "1"?

Thanks,

The loop iteration space is normalized.

Best regards,
Alexey Bataev

Thanks. Is there a way to get the actual loop increment. I am comparing two loops wrt the increment value. It seems I always get the same “1” value no matter how I define the loop increment.

At LLVM IR level - no, only in the AST.

Best regards,
Alexey Bataev