Use compact compiled mailboxes

[SandroMaglione]

Summary

• replace Effect Queue with a compact FIFO for on-demand compiled statecharts
• retain Effect Queue unchanged for persistent custom Machine.logic processes
• preserve FIFO order, repeated idle wake-up, terminal send rejection, and stop arbitration with focused stress coverage

Changeset

• Added or updated for a library or package-metadata change
• Not required because this PR does not change src/ or package.json

Validation

• pnpm check
• Relevant example checks, when examples changed (not applicable)
• Reviewed the automated type-performance report, when the public TypeScript API or inference changed (pnpm perf:types; public API unchanged)
• Reviewed the automated runtime-performance report, when runtime behavior changed

Local paired runs show approximately 5-6% lower compiled-machine heap and 4-5% lower parent-child heap, with throughput and lifecycle effectively neutral to slightly positive.

[github-actions]

Type performance

Measured with TypeScript 6.0.3 and skipLibCheck=true.

Scenario	Base	PR	Difference
Effect only	55	55	0 (0.0%)
Import effect-machine	55	55	0 (0.0%)
Machine.defineStates (3 states)	2,965	2,965	0 (0.0%)
Machine.make (3 states, 2 events)	8,715	8,715	0 (0.0%)
machine.handle (3 states, 2 transitions)	23,492	23,492	0 (0.0%)
machine.handle (depth 8)	117,075	117,075	0 (0.0%)
machine.handle (depth 12)	129,713	129,713	0 (0.0%)
machine.handle (depth 16)	145,055	145,055	0 (0.0%)
machine.handle (depth 24)	183,851	183,851	0 (0.0%)
machine.handle (wide depth 16)	214,189	214,189	0 (0.0%)
machine.handle (parallel/history/choice)	132,667	132,667	0 (0.0%)
machine.handle (4 successive calls)	128,674	128,674	0 (0.0%)
machine exact input/output/error/services	110,584	110,584	0 (0.0%)
execution adapter readiness	127,854	127,854	0 (0.0%)

Marginal instantiations are measured against the matching setup without that API call:

Scenario	Base	PR	Difference
Import effect-machine	0	0	0
Machine.defineStates (3 states)	2,910	2,910	0 (0.0%)
Machine.make (3 states, 2 events)	5,742	5,742	0 (0.0%)
machine.handle (3 states, 2 transitions)	14,777	14,777	0 (0.0%)
machine.handle (depth 8)	109,601	109,601	0 (0.0%)
machine.handle (depth 12)	121,039	121,039	0 (0.0%)
machine.handle (depth 16)	135,181	135,181	0 (0.0%)
machine.handle (depth 24)	171,577	171,577	0 (0.0%)
machine.handle (wide depth 16)	202,224	202,224	0 (0.0%)
machine.handle (parallel/history/choice)	114,331	114,331	0 (0.0%)
machine.handle (4 successive calls)	113,703	113,703	0 (0.0%)
machine exact input/output/error/services	100,670	100,670	0 (0.0%)
execution adapter readiness	102,092	102,092	0 (0.0%)

Check times (informational)

Scenario	Base	PR
Effect only	0.03s	0.03s
Import effect-machine	0.03s	0.03s
Machine.defineStates (3 states)	0.11s	0.11s
Machine.make (3 states, 2 events)	0.17s	0.17s
machine.handle (3 states, 2 transitions)	0.24s	0.23s
machine.handle (depth 8)	0.56s	0.54s
machine.handle (depth 12)	0.58s	0.59s
machine.handle (depth 16)	0.63s	0.63s
machine.handle (depth 24)	0.74s	0.74s
machine.handle (wide depth 16)	0.79s	0.81s
machine.handle (parallel/history/choice)	0.65s	0.66s
machine.handle (4 successive calls)	0.60s	0.61s
machine exact input/output/error/services	0.53s	0.54s
execution adapter readiness	0.59s	0.61s

Type instantiations are the comparison metric. Check time varies with runner load and is informational only.

[github-actions]

Runtime performance

Median of 5 independent benchmark processes on AMD EPYC 7763 64-Core Processor with Node v24.18.0.

Pull request baseline

Scenario	Effect Machine	XState 5	XState 6 alpha
Plan counter transitions	100,701 transitions/s	136,436 transitions/s	123,960 transitions/s
Drain burst with terminal fence	82,879 increments/s	404,742 increments/s	112,903 increments/s
Drain burst with a change observer	61,125 increments/s	403,061 increments/s	112,416 increments/s
Lookup and send to one child	76,451 increments/s	399,794 increments/s	107,947 increments/s
Start and stop a machine	17,013 machines/s	184,502 machines/s	72,696 machines/s
Start and stop a parent with one child	5,587 families/s	67,395 families/s	27,020 families/s

Effect runtime reference points

Scenario	Effect Machine	Effect runtime primitives
Start and stop a raw compiled process	24,398 processes/s	—
Start and interrupt a suspended fiber	—	164,177 fibers/s
Start and stop a queue worker	—	113,688 workers/s
Start and stop an actor shell	—	104,297 actors/s
Start and stop two actor shells	—	58,268 families/s
Update an owner-only mutable snapshot	—	317,863,954 updates/s
Update a synchronized snapshot	—	1,435,925 updates/s
Create, resolve, and await a terminal latch	—	1,408,010 latches/s

Memory profile	Effect Machine	XState 5	XState 6 alpha	Effect runtime primitives
Idle machine	8.4 KiB	3.8 KiB	6.1 KiB	—
Raw generic managed process	13.2 KiB	—	—	—
Raw compiled process	9.1 KiB	—	—	—
Two independent idle machines	16.0 KiB	7.3 KiB	11.8 KiB	—
Idle parent with one child	21.1 KiB	5.5 KiB	8.8 KiB	—
Parent with observed child registry	32.5 KiB	—	—	—
Parent with observed invoked child snapshots	21.3 KiB	—	—	—
Suspended Effect fiber	—	—	—	0.6 KiB
Effect queue with waiting fiber	—	—	—	2.8 KiB
Effect mailbox actor shell	—	—	—	3.2 KiB
Two Effect actor shells	—	—	—	6.9 KiB

Effect Machine change from base

Metric	Base	Base variability	PR	PR variability	Difference
Plan counter transitions	100,056 transitions/s	3.2% MAD	100,701 transitions/s	0.5% MAD	+0.6%
Drain burst with terminal fence	81,499 increments/s	3.1% MAD	82,879 increments/s	1.0% MAD	+1.7%
Drain burst with a change observer	61,546 increments/s	3.6% MAD	61,125 increments/s	1.8% MAD	-0.7%
Lookup and send to one child	74,506 increments/s	3.0% MAD	76,451 increments/s	1.7% MAD	+2.6%
Start and stop a machine	16,299 machines/s	2.2% MAD	17,013 machines/s	2.8% MAD	+4.4%
Start and stop a parent with one child	5,266 families/s	3.3% MAD	5,587 families/s	1.6% MAD	+6.1%
Idle machine heap per unit	8.8 KiB	0.0% MAD	8.4 KiB	0.0% MAD	-5.2%
Raw generic managed process heap per unit	13.2 KiB	0.0% MAD	13.2 KiB	0.0% MAD	+0.0%
Raw compiled process heap per unit	9.6 KiB	0.0% MAD	9.1 KiB	0.1% MAD	-5.5%
Two independent idle machines heap per unit	17.1 KiB	0.0% MAD	16.0 KiB	0.0% MAD	-6.2%
Idle parent with one child heap per unit	22.1 KiB	0.0% MAD	21.1 KiB	0.0% MAD	-4.7%
Parent with observed child registry heap per unit	33.6 KiB	0.0% MAD	32.5 KiB	0.0% MAD	-3.3%
Parent with observed invoked child snapshots heap per unit	22.3 KiB	0.0% MAD	21.3 KiB	0.0% MAD	-4.5%

Effect runtime reference change from base

Metric	Base	Base variability	PR	PR variability	Difference
Start and stop a raw compiled process	23,970 processes/s	3.1% MAD	24,398 processes/s	1.0% MAD	+1.8%
Start and interrupt a suspended fiber	164,989 fibers/s	0.8% MAD	164,177 fibers/s	1.6% MAD	-0.5%
Start and stop a queue worker	114,469 workers/s	0.4% MAD	113,688 workers/s	0.5% MAD	-0.7%
Start and stop an actor shell	105,955 actors/s	1.4% MAD	104,297 actors/s	1.4% MAD	-1.6%
Start and stop two actor shells	65,928 families/s	1.8% MAD	58,268 families/s	1.4% MAD	-11.6%
Update an owner-only mutable snapshot	316,957,211 updates/s	0.0% MAD	317,863,954 updates/s	0.0% MAD	+0.3%
Update a synchronized snapshot	1,500,353 updates/s	1.0% MAD	1,435,925 updates/s	0.9% MAD	-4.3%
Create, resolve, and await a terminal latch	1,469,281 latches/s	0.1% MAD	1,408,010 latches/s	1.0% MAD	-4.2%

Versions and interpretation

• Effect Machine: 0.3.0
• XState 5: 5.32.5
• XState 6 alpha: 6.0.0-alpha.27
• Effect runtime primitives: 4.0.0-beta.105

Higher throughput is better; lower heap is better. Variability is the median absolute deviation across independent processes, relative to their median. Runtime measurements on shared GitHub-hosted hardware remain informational, so small differences should be confirmed across multiple workflow runs.