npub1pssyplhacfavrncaa572yn2ac4pcjruy79u58j0xmmxpyel9vh2s4p6x53 (npub1pss…6x53) yes, if you tune - Xmx/-Xms and friends properly. And if every corner of your code knows how to deal with complex hierarchies of objects, knowing that a forgotten reference is forever.
In practice, I run several Java applications on my prod servers (Elastic, Kafka, Keycloak...), and even with very low workloads they end up taking between 500M-1G of memory - or whatever you allocate for -Xmx.
If a language's memory performance requires both developers to know how garbage collection works and how they should handle references (because, unlike Rust, the compiler won't tell you anything if you leak references across contexts), and sysadmins to tweak heap allocation options based on heuristic estimates of how much memory the application requires, then that language has a long way to go towards sustainable memory management.
Fabio Manganiello /
npub13u…dpvgs
2024-01-28 22:26:47
in reply to nevent1q…dlx2
Author Public Key
npub13uunvh7djw9ep54nswkuxlneyee7ehcpc7e53t68krykrdeg6j4qrdpvgsPublished at
2024-01-28 22:26:47Event JSON
{
"id": "36f10b094e4e76040b696c06a0d65fa666fa40722a351706bd280f9e6d03142a",
"pubkey": "8f39365fcd938b90d2b383adc37e792673ecdf01c7b348af47b0c961b728d4aa",
"created_at": 1706480807,
"kind": 1,
"tags": [
[
"p",
"0c2040fefdc27ac1cf1ded3ca24d5dc543890f84f17943c9e6decc1267e565d5",
"wss://relay.mostr.pub"
],
[
"p",
"e199692de4124b2c4a9e06c5d6affe667d4a69ca9f3df37ef8ca7c3913265f0e",
"wss://relay.mostr.pub"
],
[
"e",
"c83569c84e61130166a366ac786f2ef62905d8cb34ad4f86a8feb5f58af08269",
"wss://relay.mostr.pub",
"reply"
],
[
"proxy",
"https://manganiello.social/objects/b35f47e6-72e1-4712-8b0c-f3d5d2ffb9f2",
"activitypub"
]
],
"content": "nostr:npub1pssyplhacfavrncaa572yn2ac4pcjruy79u58j0xmmxpyel9vh2s4p6x53 yes, if you tune - Xmx/-Xms and friends properly. And if every corner of your code knows how to deal with complex hierarchies of objects, knowing that a forgotten reference is forever.\n\nIn practice, I run several Java applications on my prod servers (Elastic, Kafka, Keycloak...), and even with very low workloads they end up taking between 500M-1G of memory - or whatever you allocate for -Xmx.\n\nIf a language's memory performance requires both developers to know how garbage collection works and how they should handle references (because, unlike Rust, the compiler won't tell you anything if you leak references across contexts), and sysadmins to tweak heap allocation options based on heuristic estimates of how much memory the application requires, then that language has a long way to go towards sustainable memory management.",
"sig": "9edbecb11b62ed27b3b0e971daaf8a6d006d65ff9937136050f9af99d6895e63e6b0ba6d4e7769fc6028e201ee972cef369bedaada27fb0909674d925895f046"
}