what a wonderful question! Since you didn't include a move number, I can only work with the various positions where black's dark-squared bishop had moved to g5, or where it was on g5 for several moves.

It first went to on black's fifteenth move, following which white played 16. Ne3. Had black exchanged the bishop for the knight in that position, forcing white to capture with 17. Fe3, the resultant pawn formation would have been bad for black because of his fianchettoed king's position with pawns at g6 and f6 and no dark-squared bishop to plug up the holds. that is before you take into account the half-open f-file and open hole on f6.

In addition to this, in that line, white's pawn on e3 would have kept black's knight out of the d4 square while white's pieces had a perfect outpost on d5 which could have been supported by a later e3-e4 push, so that exchanging off whatever piece white posted at d5 would finally have been replaced by a protected passed pawn that black would have had to either blockade or undermine in order to try to maintain the balance in an endgame.

Black's bishop's second vacation trip to g5 was 29... Bg5. At this point, it was part of black's attempt at counter-attack, and in most of the subsequent moves, it was burdened with protecting the rook on h4 after 30...Rh4 which is what forced the exchange of queens (the B on g5 was en prise) and gave white a strong initiative after (33. rf5 be7, 34 d6 Bd8, 35. g5 rh5 36. Kg4...) which allowed white to steamroller black in short order.