prove that we actually benefit from `jemalloc`

0xPolygonZero / zk_evm

Apache License 2.0

69 stars 21 forks source link

prove that we actually benefit from `jemalloc` #302

Open 0xaatif opened 3 weeks ago

0xaatif commented 3 weeks ago

FWIW I would normally want to see profiling before opting-in to any allocator change, but I don't want to get bogged down in that discussion for now

_Originally posted by @0xaatif in https://github.com/0xPolygonZero/zk_evm/issues/301#issuecomment-2178438342_

BGluth commented 2 weeks ago

FWIW, @nbgl (hardcore optimization engineer) mentioned that we might be getting a non-negligible benefit from jemalloc at this point. I don't know how much hardware influences the performance of Jemalloc, but if it's significant, that might make benchmarking a bit more interesting.