A* Search G Score calculation.

[goldenchicken]

I am trying to implement the A* search algorithm, and I have the follwing function to calculate the G scores for every node. Node that I assume a normal grid as the node graph. My problem is that the G score calculation is quite slow. Is there any more efficient way of doing this?

	void AStarSearch::calculateGScore(int x, int y, int prevScore, std::vector<std::vector<bool> > &visitedGScore)
	{
		/*Get new score*/
		float newScore = prevScore + 1;
		/*If new score is greater than old score, and we have already visited the node, return*/
		if (visitedGScore[y][x] == true && nodes[y][x].getG() < newScore)
			return;

		nodes[y][x].setG(newScore);

		/*Set visited*/
		visitedGScore[y][x] = true;

		/*Recursively call calculateGScore for surrounding nodes
		  We need to test to make sure we don't go out of bounds. We also have to check if the node is walkable.
		  Note that we cant' go out of bounds in the if statement because the second condition won't be tested if the first
		  is false.*/
		if (x > 0 && walkableNodes[y][x - 1])
			calculateGScore(x - 1, y, newScore, visitedGScore);
		if (x < width - 1 && walkableNodes[y][x + 1])
			calculateGScore(x + 1, y, newScore, visitedGScore);
		if (y > 0 && walkableNodes[y - 1][x])
			calculateGScore(x, y - 1, newScore, visitedGScore);
		if (y < height - 1 && walkableNodes[y + 1][x])
			calculateGScore(x, y + 1, newScore, visitedGScore);
	}