You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
pbkit currently has very bad performance at encoding/decoding messages (~20 times worse than protobufjs)
pbkit encode x 41,223 ops/sec ±0.35% (93 runs sampled)
protobufjs encode x 843,774 ops/sec ±0.51% (93 runs sampled)
pbkit decode x 36,165 ops/sec ±0.94% (91 runs sampled)
protobufjs decode x 281,256 ops/sec ±0.40% (93 runs sampled)
It seems obvious that the Long constructor (dark-blue part in below screenshot), which does a lot of unnecessary work, is the bottleneck most of the time.
(left: pbkit, right: protobufjs)
(Bottom up chart)
You can check detail from this cpu profile snapshots:
Moved from #112
pbkit currently has very bad performance at encoding/decoding messages (~20 times worse than protobufjs)
It seems obvious that the Long constructor (dark-blue part in below screenshot), which does a lot of unnecessary work, is the bottleneck most of the time.
(left: pbkit, right: protobufjs)
(Bottom up chart)
You can check detail from this cpu profile snapshots:
pbkit-benchmarks-cpuprofiles.zip
The text was updated successfully, but these errors were encountered: