
Is what you see when you do an allocation profile of kubo when under load (ipfs add & ipfs get & gateway).
The goal is to reduce CPU usage by running GC less often, reduce stop the world events (and cooperative preamptions) and finaly give better latency.
This is a collection of smallish refactors that we should fix :
Is what you see when you do an allocation profile of kubo when under load (
ipfs add&ipfs get& gateway).The goal is to reduce CPU usage by running GC less often, reduce stop the world events (and cooperative preamptions) and finaly give better latency.
This is a collection of smallish refactors that we should fix :
blocks.Blockgo-block-format#45multiaddr.Protocols, adding amultiaddr.AppendProtocols(and using it), woud allow to reuse slice capacity: perf: Add AppendProtocols for a an allocation free to get the protocols multiformats/go-multiaddr#188