lnd.conf file tweaking for high-performance hardware

[SunnySarahNode]

There is a lot of information on how to optimize the LND client for weak hardware like Raspberry Pi, but there is nothing on how to optimize medium or large nodes for high performance.

Let's assume you have a good server with a powerful CPU, 128 Gb of RAM or more, fast storage units and so on. What would your lnd.conf file settings be to take full advantage of this configuration?

I'm guessing something like (conservative) :

blockcachesize=1000000000 -- ~1Gb of RAM, connect to backend less often
gossip.msg-rate-bytes=1000000 -- for not collecting gossips in a queue
gossip.msg-burst-bytes=2000000 -- same
gossip.filter-concurrency=10 -- or even more?

caches.reject-cache-size=100000000 -- ~2.5Gb of RAM for speeding up channel updates filtration
caches.channel-cache-size=40000000 -- ~8Gb of RAM for speeding up the response to other nodes

Thinking of

db.batch-commit-interval=2500ms -- (or more) increase RAM usage, decrease disk i/o, but not sure which backends this applies to and how safe/useful is it.

Will there be any benefit from this? What else can I pay attention to?

Glad to hear your thoughts on this.