Performance report — pg.Object vs dataclass / pydantic

Active core: rust (native pygx_core when rust). Median ns/op (testing/benchmarks/_bench.py auto-scaling timer); lower is better; — = no equivalent. Numbers are machine-specific (Apple Silicon here) — read ratios, not absolutes.

Regenerate: PYGX_CORE=rust .venv-coredev/bin/python testing/benchmarks/perf_report_bench.py

Notes: 8 fields + 4 kwargs adds 4 undeclared kwargs at construction. pygx Objects (topo=False and topo=True) and dataclass are typed/closed and reject undeclared kwargs by design, so construct / from-json show —; only pydantic (extra="allow") absorbs them. (Read-only ops for that shape run on the 8 declared fields.) symbolic children nests a 2-field object as a field. to/from-json is not identical work across columns: pygx emits a _type tag for round-trippable typed reconstruction (and dispatches on it when loading), whereas dataclass asdict / pydantic model_dump emit bare dicts — extra work attributed to pygx. Class-level utilities (register/load/save/...) and on_sym_* hooks are excluded (not per-instance ops).

Section A — all public operations (3 scalar fields)

Operation	topo=False	topo=True	dataclass	pydantic
Lifecycle (dataclass / pydantic analog)
construct (kwargs)	292.4 ns	282.7 ns	191.8 ns	522.3 ns
attr get	44.6 ns	44.5 ns	39.7 ns	45.4 ns
attr set	139.7 ns	138.6 ns	52.1 ns	142.6 ns
eq (==, same)	312.4 ns	391.3 ns	130.2 ns	288.3 ns
hash (frozen)	459.0 ns	445.3 ns	101.5 ns	113.2 ns
repr	558.8 ns	555.4 ns	306.3 ns	1.03 us
str	564.2 ns	560.5 ns	303.9 ns	909.9 ns
clone (shallow)	490.8 ns	515.7 ns	624.9 ns	592.0 ns
clone (deep)	914.6 ns	881.6 ns	2.27 us	1.62 us
to dict/json	388.6 ns	387.9 ns	631.0 ns	531.8 ns
to json str	2.22 us	2.19 us	1.51 us	596.8 ns
from dict/json	977.4 ns	981.9 ns	195.6 ns	656.2 ns
from json str	3.43 us	3.46 us	—	532.5 ns
Symbolic sym_* (no analog)
sym_clone	511.6 ns	528.2 ns	—	—
sym_rebind	887.2 ns	869.9 ns	—	—
sym_seal / unseal	91.3 ns	90.7 ns	—	—
sym_eq	80.7 ns	81.2 ns	—	—
sym_ne	110.6 ns	110.2 ns	—	—
sym_lt	1.20 us	1.22 us	—	—
sym_gt	1.43 us	1.46 us	—	—
sym_hash	403.1 ns	399.0 ns	—	—
sym_missing	566.2 ns	566.3 ns	—	—
sym_nondefault	2.14 us	2.13 us	—	—
sym_keys	226.7 ns	228.4 ns	—	—
sym_values	226.3 ns	226.3 ns	—	—
sym_items	248.6 ns	248.3 ns	—	—
sym_hasattr	168.3 ns	167.6 ns	—	—
sym_getattr	187.4 ns	186.8 ns	—	—
sym_get	445.5 ns	454.5 ns	—	—
sym_has	272.7 ns	275.8 ns	—	—
sym_contains	502.1 ns	494.8 ns	—	—
sym_jsonify	1.03 us	1.30 us	—	—
sym_descendants	1.10 us	1.11 us	—	—
sym_attr_field	172.5 ns	171.9 ns	—	—
format(compact)	403.0 ns	404.2 ns	—	—
query (pg.query)	5.99 us	6.07 us	—	—
Properties sym_* (no analog)
sym_partial	297.1 ns	298.7 ns	—	—
sym_abstract	495.3 ns	504.2 ns	—	—
sym_puresymbolic	178.8 ns	180.4 ns	—	—
sym_sealed	98.0 ns	97.9 ns	—	—
topo_root	—	197.5 ns	—	—
sym_origin	97.9 ns	97.6 ns	—	—
sym_init_args	1.14 us	1.19 us	—	—
topo_path	—	90.2 ns	—	—
topo_parent	—	90.7 ns	—	—
enable_topo	49.7 ns	49.7 ns	—	—
accessor_writable	97.5 ns	97.8 ns	—	—
allow_partial	97.8 ns	97.6 ns	—	—
is_deterministic	1.31 us	1.31 us	—	—

Section B — key operations across scales

construct

Shape	topo=False	topo=True	dataclass	pydantic
3 fields	289.6 ns	290.0 ns	201.6 ns	534.5 ns
4 fields	337.7 ns	335.0 ns	248.0 ns	593.1 ns
8 fields	512.7 ns	508.4 ns	528.9 ns	899.5 ns
16 fields	849.8 ns	860.7 ns	1.27 us	1.49 us
8 fields + 4 kwargs	—	—	—	1.22 us
container children	756.3 ns	821.9 ns	244.9 ns	785.0 ns
symbolic children	1.22 us	2.97 us	320.0 ns	853.7 ns

attr get

Shape	topo=False	topo=True	dataclass	pydantic
3 fields	39.0 ns	39.1 ns	34.9 ns	40.3 ns
4 fields	40.0 ns	40.3 ns	35.0 ns	41.1 ns
8 fields	39.4 ns	39.9 ns	34.8 ns	40.5 ns
16 fields	40.6 ns	40.0 ns	34.1 ns	41.5 ns
8 fields + 4 kwargs	40.8 ns	40.7 ns	35.4 ns	40.8 ns
container children	40.2 ns	39.6 ns	35.3 ns	40.8 ns
symbolic children	40.1 ns	41.0 ns	35.8 ns	42.1 ns

eq

Shape	topo=False	topo=True	dataclass	pydantic
3 fields	302.0 ns	384.7 ns	119.8 ns	283.9 ns
4 fields	315.2 ns	431.5 ns	141.0 ns	292.4 ns
8 fields	342.9 ns	597.0 ns	232.3 ns	316.1 ns
16 fields	387.0 ns	938.3 ns	429.7 ns	363.5 ns
8 fields + 4 kwargs	349.0 ns	607.1 ns	240.8 ns	323.5 ns
container children	536.1 ns	539.0 ns	176.9 ns	328.5 ns
symbolic children	949.2 ns	975.3 ns	220.3 ns	607.2 ns

clone

Shape	topo=False	topo=True	dataclass	pydantic
3 fields	488.6 ns	520.0 ns	634.2 ns	595.5 ns
4 fields	507.0 ns	548.8 ns	664.5 ns	602.7 ns
8 fields	542.5 ns	616.0 ns	679.9 ns	645.5 ns
16 fields	591.2 ns	724.4 ns	802.8 ns	748.5 ns
8 fields + 4 kwargs	552.6 ns	619.7 ns	711.6 ns	641.5 ns
container children	545.5 ns	571.2 ns	686.2 ns	647.6 ns
symbolic children	533.1 ns	3.78 us	676.3 ns	617.2 ns

clone/deep

Shape	topo=False	topo=True	dataclass	pydantic
3 fields	888.9 ns	884.7 ns	2.20 us	1.64 us
4 fields	971.4 ns	971.0 ns	2.51 us	1.88 us
8 fields	1.31 us	1.31 us	3.52 us	2.84 us
16 fields	1.98 us	1.97 us	5.47 us	4.77 us
8 fields + 4 kwargs	1.37 us	1.35 us	3.50 us	3.20 us
container children	4.24 us	4.30 us	4.58 us	3.92 us
symbolic children	2.46 us	3.88 us	4.55 us	3.43 us

to json

Shape	topo=False	topo=True	dataclass	pydantic
3 fields	391.9 ns	394.4 ns	647.3 ns	530.0 ns
4 fields	415.7 ns	429.8 ns	752.1 ns	571.3 ns
8 fields	538.9 ns	638.3 ns	1.17 us	730.3 ns
16 fields	770.3 ns	1.05 us	2.08 us	1.07 us
8 fields + 4 kwargs	541.4 ns	639.7 ns	1.19 us	768.5 ns
container children	668.8 ns	10.17 us	1.74 us	827.6 ns
symbolic children	2.30 us	647.2 ns	1.22 us	708.0 ns

from json

Shape	topo=False	topo=True	dataclass	pydantic
3 fields	978.4 ns	969.8 ns	207.7 ns	648.2 ns
4 fields	1.12 us	1.13 us	256.1 ns	709.5 ns
8 fields	1.61 us	1.59 us	537.5 ns	987.4 ns
16 fields	2.48 us	2.49 us	1.30 us	1.43 us
8 fields + 4 kwargs	—	—	—	1.18 us
container children	1.77 us	1.86 us	244.9 ns	924.6 ns
symbolic children	1.98 us	3.66 us	310.7 ns	968.2 ns