It is obvious that in the current pawn battle, White has the better chances. However, after an exchange of one pair of pawns on the queenside and the appearance there of the black knight, the win would be problematical. White’s solution is elegant and convincing:

The pawn on b6 is defended, the Rb7 passive, and the only other blockader of the white pawns is eliminated on g4. There is no way to stop the white pawns, so Black resigned.

The motif of advancing the pawns closer to queen strikes one at first glance. And certainly, with his knight so far away, the white queen has trouble coping with the excellently-coordinated black pieces, supporting their passed pawn.

30...Qxg2! Qxg2! 31.Rg1 f3 32.Nb5 Nxb5 33.Rxg2+ fxg2 34.cxb5?

Now, the black bishop can support the passed pawn from f1 rather than h3, which facilitates the bringing up of additional help in the form of the h-pawn.
It was essential to play 34.Qg1! Bh3 (34...Nd4 35.Qxg2+ Kf7 36.f3 h5 leads to something close to a positional draw –

White resigned, since after 43.Kd3, there follows 43...c4+.

The white e-pawn cannot advance, because the d4-pawn would be hanging. This means that it needs to be blockaded, so as to leave White with a bad bishop. Obviously, to do this, we need the knight. The route to e4 is obvious:

The attack on the rook is extremely useful for Black, as it effectively shortens by one tempo the route to e4.

Definitely the right move, forcing the exchange of bishops. The superiority of the knight over the remaining white bishop will be overwhelming.

This desperate attempt at counterplay does not give any chances – the advance of the pawn is not supported by the white pieces. Black reacts in the simplest possible way, keeping the position closed, which suits him.

The queen alone cannot give mate, but it was hard to recommend anything to White – Black is already prepared to break in down the c-file.