I would no longer use nodes as a yardstick for measuring the quality of a chess engine. But the longer the game, the more positional it gets. Even today, Deep Shredder 13 manages to hold it's own in longer time controls. -It's not really competing with the top 5 engines in any time control anymore, but it has always done much better at the slower time controls. And I would still argue that for analyzing games at a longer time control, it's still a good engine to invest in.
There really is no "best" engine for analyzing games. Every engine has it's own algorithm and brings to the table it's own set of insights into a game or position your analyzing.
This graph tell you general idea of how strong your stockfish 10+ developmental version (may 2019 )will be when you give a certain amount of time.
In general, 4 cores cpu get average 5 millions nodes per second.
0.1 seconds of analysis per position will be 3200 rating.
1 sec per position of analysis will be 3500 rating.
1 min per position of analysis will be 3700 rating,
but the elo strength become flat, so even if you do for 1 day analysis, rating may be just minimally increase.
However, in theory if you do instant move or almost 0 seconds for analysis for stockfish, stockfish rating may be as low as 900 rating.
Note, Rating are compared to fast GM rating .http://fastgm.de/