Benchmarking postgres.js vs pg vs pg-native

[nigrosimone]

Hi,

I created a benchmark to compare the performance between postgres.js, pg, and pg-native: https://github.com/nigrosimone/postgres-benchmarks

Results

Running benchmarks... --iterations=50000
GC is exposed
{
  "pg": "^8.16.3",
  "pg-native": "^3.5.2",
  "postgres": "^3.4.7"
}
Database connectivity verified
clk: ~3.23 GHz
cpu: Intel(R) Core(TM) i7-1065G7 CPU @ 1.30GHz
runtime: node 24.7.0 (x64-linux)

benchmark                       avg (min … max) p75 / p99    (min … top 1%)
----------------------------------------------- -------------------------------
brianc/node-postgres (pg-native) 177.55 µs/iter 187.29 µs  █
                          (115.42 µs … 2.16 ms) 499.22 µs  █
                        (  5.93 kb …   2.35 mb)  12.06 kb ▁█▇▄▃▃▃▃▂▂▂▁▁▁▁▁▁▁▁▁▁

brianc/node-postgres (pg)        200.38 µs/iter 224.83 µs   █
                          (117.27 µs … 1.98 ms) 492.05 µs  ▅█
                        (  2.37 kb …   1.82 mb)  15.24 kb ▁███▇▇▆▅▄▂▂▂▂▂▁▁▁▁▁▁▁

porsager/postgres (postgres)     196.89 µs/iter 203.13 µs   █
                          (120.43 µs … 1.87 ms) 533.71 µs   █
                        (  7.00 kb …   1.18 mb)  12.01 kb ▁▇██▅▄▃▂▂▂▁▁▁▁▁▁▁▁▁▁▁

summary
  brianc/node-postgres (pg-native)
   1.11x faster than porsager/postgres (postgres)
   1.13x faster than brianc/node-postgres (pg)

From these results, pg-native appears to be slightly faster than both pg and postgres.js.

Is my benchmark methodology correct?
Is there something I'm missing that could explain these performance differences?

Thanks!

[mattbishop]

It doesn’t seem to do much. I’d like to see some more significant data flows—how well do they process inserts and select results?—Matt BishopOn Sep 6, 2025, at 12:21 PM, Nigro Simone ***@***.***> wrote: Hi, I created a benchmark to compare the performance between postgres.js, pg, and pg-native: https://github.com/nigrosimone/postgres-benchmarks Results Running benchmarks... --iterations=50000 GC is exposed { "pg": "^8.16.3", "pg-native": "^3.5.2", "postgres": "^3.4.7" } Database connectivity verified clk: ~3.23 GHz cpu: Intel(R) Core(TM) i7-1065G7 CPU @ 1.30GHz runtime: node 24.7.0 (x64-linux) benchmark avg (min … max) p75 / p99 (min … top 1%) ----------------------------------------------- ------------------------------- brianc/node-postgres (pg-native) 177.55 µs/iter 187.29 µs █ (115.42 µs … 2.16 ms) 499.22 µs █ ( 5.93 kb … 2.35 mb) 12.06 kb ▁█▇▄▃▃▃▃▂▂▂▁▁▁▁▁▁▁▁▁▁ brianc/node-postgres (pg) 200.38 µs/iter 224.83 µs █ (117.27 µs … 1.98 ms) 492.05 µs ▅█ ( 2.37 kb … 1.82 mb) 15.24 kb ▁███▇▇▆▅▄▂▂▂▂▂▁▁▁▁▁▁▁ porsager/postgres (postgres) 196.89 µs/iter 203.13 µs █ (120.43 µs … 1.87 ms) 533.71 µs █ ( 7.00 kb … 1.18 mb) 12.01 kb ▁▇██▅▄▃▂▂▂▁▁▁▁▁▁▁▁▁▁▁ summary brianc/node-postgres (pg-native) 1.11x faster than porsager/postgres (postgres) 1.13x faster than brianc/node-postgres (pg) From these results, pg-native appears to be slightly faster than both pg and postgres.js. Is my benchmark methodology correct? Is there something I'm missing that could explain these performance differences? Thanks! —Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>