Gregory Popovitch

Looks great and very simple indeed. I don't know enough about spinlocks to comment on using it instead of std::mutex.

> the new `packed_spinlock` code, which should (in theory) help avoid contention for the cache lock under most scenarios. Just a thought. Maybe we never need to wait? We could...

Thanks @scottschurr, very useful. Some random thoughts I have. 1. strange that `longest ns` is longer for Scott D's branch than `develop`. I would assume these occur when we don't...

Libération ?

@yperbasis would it be possible to implement SilkRPC in a separate process (accessing the same libmdbx database) to avoid the async API contagion?

@AndreaLanfranchi , does Erigon's rpcdaemon use asynchronous i/o? Does it help with performance of the rpc endpoint?

> * shared memory : rpcdaemon process opens the same MDBX environment as Erigon's in shared mode (so they share the same memory mapping). Better performance > * remote :...

My idea was to execute the last 1000 blocks of the ethereum blockchain, while tracking any accessed mdbx keys, and their value the first time they are accessed. This would...

@chfast, let me know what you find out please :-)

Thanks, glad to hear :-)