Patch: Reorganize core Networking Structs to optimize cacheline consumption (2023) (lore.kernel.org) 4 points by hyperpape 2y ago ↗ HN
[–] jauntywundrkind 2y ago ↗ "Epic" wins from re-ordering some structs for cache efficiency. AMD seeing a 20-40% win across the board.Interestingly a much more mild impact on Intel systems; makes me think they're doing something really right with predictive fetch somewhere.So excellent having such shared communal property everyone is investing in. Great stuff here.
2 comments
[ 4.6 ms ] story [ 15.1 ms ] threadInterestingly a much more mild impact on Intel systems; makes me think they're doing something really right with predictive fetch somewhere.
So excellent having such shared communal property everyone is investing in. Great stuff here.