Eliminate redundant OP_MOVE in plain function call compilation

[frostney]

Summary

• Optimizes plain function call compilation to eliminate redundant `OP_MOVE` instructions.
• When `ADest` is at the top of the register stack and not a local variable slot, the call is compiled directly into `ADest` instead of a temp register.
• Fires for expression statements, binary expression operands, nested call arguments, and other contexts where a fresh temp was just allocated.
• Method calls, super calls, and private member calls are left unchanged (different register layout makes this unsafe).

Test plan

• All JS tests pass in interpreter mode
• All JS tests pass in bytecode mode

Made with Cursor

Summary by CodeRabbit

• Refactor
  • Improved register allocation and memory management during function call compilation for better efficiency.
  • Optimized conditional logic for register assignment to reduce unnecessary register operations.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: e2377881-ad43-4bd2-bf53-25f51926264d

📥 Commits

Reviewing files that changed from the base of the PR and between 7001181 and 1a6b8f4.

📒 Files selected for processing (1)

• units/Goccia.Compiler.Expressions.pas

---

📝 Walkthrough

Walkthrough

Optimizes register allocation in the generic call path of CompileCall by introducing an IsLocalSlot helper function and adding conditional logic to prefer using ADest as BaseReg when appropriate, reducing unnecessary register allocations and moves.

Changes

Cohort / File(s)	Summary
Register Allocation Optimization units/Goccia.Compiler.Expressions.pas	Added IsLocalSlot helper function and modified CompileCall to conditionally use ADest as BaseReg when ADest + 1 equals Scope.NextSlot and isn't an existing local slot. Restructured post-call register operations to conditionally move and free registers only when BaseReg differs from ADest.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~15 minutes

Poem

🐰 A register hops with graceful ease,
No extra moves to steal the breeze,
When slots align in perfect measure,
Smart allocation is our treasure! ✨

🚥 Pre-merge checks | ✅ 3

✅ Passed checks (3 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title accurately and clearly describes the main optimization: eliminating redundant OP_MOVE instructions in plain function call compilation.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment
• Commit unit tests in branch opt/dest-reg-propagation

📝 Coding Plan

• Generate coding plan for human review comments

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[github-actions]

Suite Timing

Suite	Metric	Interpreted	Bytecode
Tests	Total	3463	3463
Tests	Passed	3422 ✅	3463 ✅
Tests	Skipped	41	0
Tests	Execution	161.0ms	168.1ms
Tests	Engine	320.9ms	693.0ms
Benchmarks	Total	254	254
Benchmarks	Duration	6.48min	10.88min

_{Measured on ubuntu-latest x64.}

[github-actions]

Benchmark Results

254 benchmarks

Interpreted: 🟢 87 improved · 🔴 3 regressed · 164 unchanged · avg +4.7%
Bytecode: 🟢 49 improved · 🔴 1 regressed · 204 unchanged · avg +4.6%

arraybuffer.js — Interp: 🟢 11, 3 unch. · avg +10.1% · Bytecode: 14 unch. · avg +3.4%

Benchmark	Interpreted	Δ	Bytecode	Δ
create ArrayBuffer(0)	442,834 → 496,029	🟢 +12.0%	141,265 → 142,760	+1.1%
create ArrayBuffer(64)	434,924 → 485,601	🟢 +11.7%	137,715 → 140,465	+2.0%
create ArrayBuffer(1024)	331,038 → 388,928	🟢 +17.5%	124,577 → 127,617	+2.4%
create ArrayBuffer(8192)	141,194 → 171,656	🟢 +21.6%	78,958 → 76,815	-2.7%
slice full buffer (64 bytes)	537,777 → 579,633	🟢 +7.8%	363,643 → 383,715	+5.5%
slice half buffer (512 of 1024 bytes)	453,499 → 506,228	🟢 +11.6%	317,704 → 333,326	+4.9%
slice with negative indices	458,507 → 510,050	🟢 +11.2%	349,120 → 361,748	+3.6%
slice empty range	525,604 → 576,272	🟢 +9.6%	362,574 → 375,391	+3.5%
byteLength access	1,593,949 → 1,669,253	+4.7%	1,075,299 → 1,125,637	+4.7%
Symbol.toStringTag access	1,188,971 → 1,144,483	-3.7%	548,033 → 565,394	+3.2%
ArrayBuffer.isView	780,799 → 808,887	+3.6%	456,323 → 476,176	+4.4%
clone ArrayBuffer(64)	400,522 → 438,418	🟢 +9.5%	306,455 → 327,265	+6.8%
clone ArrayBuffer(1024)	307,778 → 354,231	🟢 +15.1%	241,145 → 249,936	+3.6%
clone ArrayBuffer inside object	273,216 → 299,350	🟢 +9.6%	150,917 → 156,999	+4.0%

arrays.js — Interp: 🟢 12, 7 unch. · avg +7.8% · Bytecode: 🟢 5, 14 unch. · avg +4.6%

Benchmark	Interpreted	Δ	Bytecode	Δ
Array.from length 100	14,336 → 15,314	+6.8%	12,844 → 12,541	-2.4%
Array.from 10 elements	234,159 → 255,762	🟢 +9.2%	148,782 → 159,735	🟢 +7.4%
Array.of 10 elements	319,137 → 344,978	🟢 +8.1%	219,792 → 223,727	+1.8%
spread into new array	348,480 → 370,286	+6.3%	568,531 → 661,103	🟢 +16.3%
map over 50 elements	28,476 → 31,006	🟢 +8.9%	22,174 → 22,235	+0.3%
filter over 50 elements	24,330 → 25,958	+6.7%	21,694 → 21,185	-2.3%
reduce sum 50 elements	27,611 → 29,388	+6.4%	19,723 → 19,478	-1.2%
forEach over 50 elements	23,411 → 24,648	+5.3%	22,664 → 23,198	+2.4%
find in 50 elements	35,726 → 38,299	🟢 +7.2%	28,180 → 27,696	-1.7%
sort 20 elements	12,416 → 13,219	+6.5%	3,180 → 3,208	+0.9%
flat nested array	117,560 → 128,995	🟢 +9.7%	292,411 → 370,778	🟢 +26.8%
flatMap	74,825 → 82,016	🟢 +9.6%	210,985 → 246,907	🟢 +17.0%
map inside map (5x5)	21,945 → 23,978	🟢 +9.3%	69,684 → 76,179	🟢 +9.3%
filter inside map (5x10)	16,417 → 17,794	🟢 +8.4%	13,543 → 13,690	+1.1%
reduce inside map (5x10)	19,812 → 21,387	🟢 +8.0%	13,852 → 14,079	+1.6%
forEach inside forEach (5x10)	16,971 → 17,733	+4.5%	15,032 → 15,572	+3.6%
find inside some (10x10)	14,051 → 15,311	🟢 +9.0%	10,716 → 10,790	+0.7%
map+filter chain nested (5x20)	5,536 → 5,989	🟢 +8.2%	4,433 → 4,513	+1.8%
reduce flatten (10x5)	39,704 → 43,605	🟢 +9.8%	5,954 → 6,177	+3.7%

async-await.js — Interp: 🟢 6 · avg +9.5% · Bytecode: 6 unch. · avg +3.2%

Benchmark	Interpreted	Δ	Bytecode	Δ
single await	381,269 → 417,744	🟢 +9.6%	268,293 → 278,695	+3.9%
multiple awaits	170,165 → 187,491	🟢 +10.2%	113,971 → 118,458	+3.9%
await non-Promise value	870,408 → 967,412	🟢 +11.1%	909,327 → 927,247	+2.0%
await with try/catch	370,025 → 407,205	🟢 +10.0%	264,985 → 271,200	+2.3%
await Promise.all	52,499 → 56,819	🟢 +8.2%	42,435 → 44,816	+5.6%
nested async function call	190,320 → 205,107	🟢 +7.8%	205,747 → 209,173	+1.7%

classes.js — Interp: 🟢 29, 2 unch. · avg +11.2% · Bytecode: 🟢 7, 24 unch. · avg +5.6%

Benchmark	Interpreted	Δ	Bytecode	Δ
simple class new	116,571 → 130,931	🟢 +12.3%	337,012 → 355,222	+5.4%
class with defaults	93,547 → 105,536	🟢 +12.8%	236,359 → 257,013	🟢 +8.7%
50 instances via Array.from	5,641 → 6,322	🟢 +12.1%	6,257 → 6,301	+0.7%
instance method call	59,503 → 67,042	🟢 +12.7%	147,565 → 162,464	🟢 +10.1%
static method call	92,175 → 102,305	🟢 +11.0%	352,299 → 376,722	+6.9%
single-level inheritance	46,362 → 52,275	🟢 +12.8%	147,419 → 162,142	🟢 +10.0%
two-level inheritance	39,573 → 44,625	🟢 +12.8%	123,586 → 138,342	🟢 +11.9%
private field access	58,425 → 65,415	🟢 +12.0%	163,573 → 178,521	🟢 +9.1%
private methods	63,715 → 71,230	🟢 +11.8%	207,363 → 229,976	🟢 +10.9%
getter/setter access	66,095 → 73,627	🟢 +11.4%	163,949 → 175,207	+6.9%
class decorator (identity)	81,089 → 91,857	🟢 +13.3%	51,978 → 53,582	+3.1%
class decorator (wrapping)	47,221 → 52,587	🟢 +11.4%	36,856 → 38,013	+3.1%
identity method decorator	58,297 → 65,572	🟢 +12.5%	42,380 → 44,813	+5.7%
wrapping method decorator	48,258 → 53,402	🟢 +10.7%	36,938 → 40,326	🟢 +9.2%
stacked method decorators (x3)	33,782 → 37,960	🟢 +12.4%	27,167 → 26,394	-2.8%
identity field decorator	65,943 → 73,383	🟢 +11.3%	44,852 → 46,695	+4.1%
field initializer decorator	56,121 → 62,435	🟢 +11.3%	40,069 → 42,008	+4.8%
getter decorator (identity)	56,459 → 63,174	🟢 +11.9%	39,258 → 40,764	+3.8%
setter decorator (identity)	47,953 → 53,244	🟢 +11.0%	34,385 → 35,758	+4.0%
static method decorator	61,819 → 69,144	🟢 +11.8%	60,474 → 64,320	+6.4%
static field decorator	71,309 → 80,117	🟢 +12.4%	64,326 → 67,726	+5.3%
private method decorator	47,730 → 52,941	🟢 +10.9%	36,375 → 37,983	+4.4%
private field decorator	52,893 → 58,521	🟢 +10.6%	37,737 → 39,195	+3.9%
plain auto-accessor (no decorator)	90,233 → 99,570	🟢 +10.3%	52,540 → 54,121	+3.0%
auto-accessor with decorator	51,206 → 58,876	🟢 +15.0%	36,623 → 37,877	+3.4%
decorator writing metadata	43,587 → 46,622	+7.0%	40,299 → 42,942	+6.6%
static getter read	104,954 → 114,818	🟢 +9.4%	377,131 → 400,964	+6.3%
static getter/setter pair	79,085 → 83,699	+5.8%	202,817 → 211,540	+4.3%
inherited static getter	59,580 → 65,897	🟢 +10.6%	260,565 → 271,240	+4.1%
inherited static setter	64,155 → 70,150	🟢 +9.3%	203,690 → 211,192	+3.7%
inherited static getter with this binding	54,330 → 58,191	🟢 +7.1%	151,331 → 161,886	+7.0%

closures.js — Interp: 11 unch. · avg +3.6% · Bytecode: 🟢 3, 8 unch. · avg +6.3%

Benchmark	Interpreted	Δ	Bytecode	Δ
closure over single variable	130,369 → 136,182	+4.5%	590,667 → 631,369	+6.9%
closure over multiple variables	121,538 → 124,547	+2.5%	370,442 → 388,979	+5.0%
nested closures	127,346 → 131,754	+3.5%	536,052 → 590,320	🟢 +10.1%
function as argument	98,824 → 100,501	+1.7%	490,382 → 529,707	🟢 +8.0%
function returning function	121,963 → 126,851	+4.0%	587,326 → 614,427	+4.6%
compose two functions	73,467 → 75,382	+2.6%	345,436 → 382,427	🟢 +10.7%
fn.call	153,902 → 159,766	+3.8%	133,133 → 136,349	+2.4%
fn.apply	114,515 → 121,902	+6.4%	89,255 → 94,931	+6.4%
fn.bind	138,189 → 146,830	+6.3%	134,106 → 141,829	+5.8%
recursive sum to 50	12,325 → 12,312	-0.1%	38,328 → 40,570	+5.9%
recursive tree traversal	19,901 → 20,817	+4.6%	61,432 → 63,414	+3.2%

collections.js — Interp: 🟢 2, 10 unch. · avg +5.4% · Bytecode: 12 unch. · avg +1.5%

Benchmark	Interpreted	Δ	Bytecode	Δ
add 50 elements	7,229 → 7,548	+4.4%	5,572 → 5,710	+2.5%
has lookup (50 elements)	90,027 → 94,669	+5.2%	84,775 → 88,347	+4.2%
delete elements	47,766 → 49,765	+4.2%	36,310 → 36,713	+1.1%
forEach iteration	16,120 → 17,004	+5.5%	15,910 → 15,986	+0.5%
spread to array	32,109 → 35,703	🟢 +11.2%	145,698 → 147,283	+1.1%
deduplicate array	41,953 → 44,632	+6.4%	45,587 → 46,991	+3.1%
set 50 entries	5,302 → 5,536	+4.4%	5,833 → 5,801	-0.6%
get lookup (50 entries)	87,782 → 89,258	+1.7%	96,294 → 97,327	+1.1%
has check	131,627 → 135,748	+3.1%	151,251 → 152,821	+1.0%
delete entries	45,578 → 46,375	+1.7%	34,899 → 35,113	+0.6%
forEach iteration	16,147 → 17,277	+7.0%	15,741 → 15,946	+1.3%
keys/values/entries	8,424 → 9,272	🟢 +10.1%	21,480 → 22,028	+2.5%

destructuring.js — Interp: 🟢 14, 8 unch. · avg +8.5% · Bytecode: 🟢 11, 11 unch. · avg +12.1%

Benchmark	Interpreted	Δ	Bytecode	Δ
simple array destructuring	428,852 → 441,984	+3.1%	620,525 → 809,899	🟢 +30.5%
with rest element	284,559 → 309,104	🟢 +8.6%	470,504 → 569,500	🟢 +21.0%
with defaults	436,999 → 443,736	+1.5%	651,382 → 787,714	🟢 +20.9%
skip elements	437,069 → 476,936	🟢 +9.1%	704,177 → 929,344	🟢 +32.0%
nested array destructuring	185,261 → 187,457	+1.2%	317,594 → 428,319	🟢 +34.9%
swap variables	532,178 → 548,879	+3.1%	933,607 → 1,110,432	🟢 +18.9%
simple object destructuring	303,125 → 327,710	🟢 +8.1%	506,853 → 522,684	+3.1%
with defaults	363,799 → 410,519	🟢 +12.8%	308,410 → 318,203	+3.2%
with renaming	315,020 → 358,590	🟢 +13.8%	591,400 → 600,160	+1.5%
nested object destructuring	147,684 → 161,389	🟢 +9.3%	245,480 → 249,940	+1.8%
rest properties	188,108 → 205,183	🟢 +9.1%	214,083 → 225,131	+5.2%
object parameter	92,007 → 102,643	🟢 +11.6%	178,589 → 182,555	+2.2%
array parameter	125,528 → 137,959	🟢 +9.9%	293,827 → 351,125	🟢 +19.5%
mixed destructuring in map	35,009 → 39,328	🟢 +12.3%	32,750 → 33,838	+3.3%
forEach with array destructuring	67,269 → 71,032	+5.6%	133,154 → 159,284	🟢 +19.6%
map with array destructuring	69,395 → 74,032	+6.7%	164,667 → 192,949	🟢 +17.2%
filter with array destructuring	72,975 → 77,679	+6.4%	195,478 → 230,125	🟢 +17.7%
reduce with array destructuring	77,513 → 82,677	+6.7%	183,069 → 208,767	🟢 +14.0%
map with object destructuring	79,842 → 87,586	🟢 +9.7%	74,024 → 73,445	-0.8%
map with nested destructuring	65,017 → 71,455	🟢 +9.9%	61,841 → 61,086	-1.2%
map with rest in destructuring	38,044 → 42,770	🟢 +12.4%	22,307 → 22,343	+0.2%
map with defaults in destructuring	58,220 → 67,024	🟢 +15.1%	37,195 → 37,687	+1.3%

fibonacci.js — Interp: 🟢 2, 6 unch. · avg +2.8% · Bytecode: 🟢 3, 5 unch. · avg +6.2%

Benchmark	Interpreted	Δ	Bytecode	Δ
recursive fib(15)	338 → 339	+0.3%	1,099 → 1,150	+4.6%
recursive fib(20)	30 → 30	+0.0%	96 → 104	🟢 +7.8%
recursive fib(15) typed	337 → 338	+0.2%	1,386 → 1,458	+5.2%
recursive fib(20) typed	31 → 30	-2.2%	118 → 132	🟢 +11.4%
iterative fib(20) via reduce	12,287 → 12,792	+4.1%	8,256 → 8,940	🟢 +8.3%
iterator fib(20)	9,464 → 9,968	+5.3%	13,731 → 14,339	+4.4%
iterator fib(20) via Iterator.from + take	14,711 → 15,842	🟢 +7.7%	15,504 → 16,413	+5.9%
iterator fib(20) last value via reduce	11,291 → 12,101	🟢 +7.2%	12,005 → 12,275	+2.2%

for-of.js — Interp: 7 unch. · avg +0.8% · Bytecode: 🟢 1, 6 unch. · avg +5.2%

Benchmark	Interpreted	Δ	Bytecode	Δ
for...of with 10-element array	47,432 → 48,242	+1.7%	147,145 → 149,769	+1.8%
for...of with 100-element array	5,562 → 5,521	-0.7%	20,920 → 21,445	+2.5%
for...of with string (10 chars)	35,377 → 35,856	+1.4%	114,503 → 121,028	+5.7%
for...of with Set (10 elements)	48,112 → 48,484	+0.8%	156,352 → 157,142	+0.5%
for...of with Map entries (10 entries)	30,920 → 31,144	+0.7%	44,393 → 50,911	🟢 +14.7%
for...of with destructuring	40,973 → 41,741	+1.9%	68,915 → 73,192	+6.2%
for-await-of with sync array	46,080 → 46,111	+0.1%	126,144 → 132,896	+5.4%

iterators.js — Interp: 🟢 2, 18 unch. · avg +2.7% · Bytecode: 🟢 5, 15 unch. · avg +3.7%

Benchmark	Interpreted	Δ	Bytecode	Δ
Iterator.from({next}).toArray() — 20 elements	14,800 → 15,849	🟢 +7.1%	15,815 → 17,003	🟢 +7.5%
Iterator.from({next}).toArray() — 50 elements	6,413 → 6,909	🟢 +7.7%	6,883 → 7,313	+6.2%
spread pre-wrapped iterator — 20 elements	11,503 → 11,827	+2.8%	15,277 → 15,252	-0.2%
Iterator.from({next}).forEach — 50 elements	4,656 → 4,846	+4.1%	4,856 → 5,296	🟢 +9.1%
Iterator.from({next}).reduce — 50 elements	4,680 → 4,844	+3.5%	4,804 → 4,987	+3.8%
wrap array iterator	177,929 → 184,921	+3.9%	102,623 → 109,680	+6.9%
wrap plain {next()} object	10,396 → 11,017	+6.0%	10,875 → 11,232	+3.3%
map + toArray (50 elements)	4,595 → 4,913	+6.9%	5,297 → 5,228	-1.3%
filter + toArray (50 elements)	4,557 → 4,701	+3.2%	5,418 → 5,134	-5.2%
take(10) + toArray (50 element source)	26,590 → 28,256	+6.3%	27,041 → 28,478	+5.3%
drop(40) + toArray (50 element source)	6,483 → 6,819	+5.2%	7,694 → 8,251	🟢 +7.2%
chained map + filter + take (100 element source)	8,682 → 8,784	+1.2%	9,161 → 9,710	+6.0%
some + every (50 elements)	2,703 → 2,779	+2.8%	3,031 → 3,253	🟢 +7.3%
find (50 elements)	5,939 → 6,007	+1.1%	6,943 → 7,118	+2.5%
array.values().map().filter().toArray()	9,420 → 9,341	-0.8%	9,315 → 9,396	+0.9%
array.values().take(5).toArray()	220,746 → 217,233	-1.6%	150,361 → 156,543	+4.1%
array.values().drop(45).toArray()	202,687 → 206,606	+1.9%	142,519 → 138,990	-2.5%
map.entries() chained helpers	11,122 → 11,090	-0.3%	4,666 → 5,232	🟢 +12.1%
set.values() chained helpers	19,224 → 19,082	-0.7%	20,284 → 19,991	-1.4%
string iterator map + toArray	14,579 → 13,751	-5.7%	20,568 → 21,147	+2.8%

json.js — Interp: 🟢 1, 19 unch. · avg +2.9% · Bytecode: 20 unch. · avg +2.4%

Benchmark	Interpreted	Δ	Bytecode	Δ
parse simple object	171,605 → 173,420	+1.1%	135,986 → 138,140	+1.6%
parse nested object	108,130 → 108,941	+0.8%	87,164 → 88,698	+1.8%
parse array of objects	57,542 → 56,815	-1.3%	50,470 → 50,262	-0.4%
parse large flat object	49,527 → 51,793	+4.6%	45,849 → 45,641	-0.5%
parse mixed types	71,774 → 74,310	+3.5%	64,107 → 67,263	+4.9%
stringify simple object	198,522 → 204,565	+3.0%	159,420 → 165,937	+4.1%
stringify nested object	113,614 → 113,840	+0.2%	87,606 → 92,523	+5.6%
stringify array of objects	62,497 → 61,592	-1.4%	51,961 → 54,075	+4.1%
stringify mixed types	87,479 → 88,883	+1.6%	72,946 → 77,331	+6.0%
reviver doubles numbers	45,719 → 47,985	+5.0%	45,151 → 45,841	+1.5%
reviver filters properties	39,219 → 41,995	🟢 +7.1%	45,554 → 46,446	+2.0%
reviver on nested object	51,992 → 54,089	+4.0%	50,921 → 52,472	+3.0%
reviver on array	29,890 → 31,353	+4.9%	28,978 → 28,771	-0.7%
replacer function doubles numbers	49,202 → 51,408	+4.5%	51,629 → 52,620	+1.9%
replacer function excludes properties	61,708 → 63,765	+3.3%	60,584 → 61,713	+1.9%
array replacer (allowlist)	117,514 → 118,895	+1.2%	97,110 → 101,706	+4.7%
stringify with 2-space indent	98,737 → 101,663	+3.0%	79,555 → 80,158	+0.8%
stringify with tab indent	98,550 → 102,119	+3.6%	79,748 → 81,924	+2.7%
parse then stringify	53,911 → 56,887	+5.5%	51,091 → 51,888	+1.6%
stringify then parse	32,479 → 33,965	+4.6%	30,694 → 30,966	+0.9%

jsx.jsx — Interp: 21 unch. · avg +2.8% · Bytecode: 21 unch. · avg +2.8%

Benchmark	Interpreted	Δ	Bytecode	Δ
simple element	208,140 → 216,514	+4.0%	613,210 → 626,176	+2.1%
self-closing element	219,486 → 225,865	+2.9%	648,972 → 666,613	+2.7%
element with string attribute	174,756 → 184,262	+5.4%	446,737 → 464,082	+3.9%
element with multiple attributes	152,012 → 153,911	+1.2%	395,627 → 398,451	+0.7%
element with expression attribute	167,104 → 175,644	+5.1%	437,654 → 446,552	+2.0%
text child	213,562 → 216,071	+1.2%	619,712 → 623,139	+0.6%
expression child	209,388 → 209,760	+0.2%	597,618 → 614,614	+2.8%
mixed text and expression	195,746 → 196,140	+0.2%	543,127 → 557,347	+2.6%
nested elements (3 levels)	78,575 → 82,797	+5.4%	242,073 → 250,737	+3.6%
sibling children	58,458 → 61,803	+5.7%	182,248 → 186,282	+2.2%
component element	148,295 → 154,639	+4.3%	418,623 → 443,935	+6.0%
component with children	91,634 → 95,842	+4.6%	269,146 → 281,385	+4.5%
dotted component	123,403 → 129,788	+5.2%	308,720 → 324,429	+5.1%
empty fragment	223,977 → 222,511	-0.7%	635,725 → 662,460	+4.2%
fragment with children	58,359 → 60,774	+4.1%	178,651 → 183,722	+2.8%
spread attributes	112,482 → 112,758	+0.2%	103,884 → 106,546	+2.6%
spread with overrides	98,708 → 99,097	+0.4%	76,900 → 78,772	+2.4%
shorthand props	160,756 → 164,524	+2.3%	409,225 → 420,191	+2.7%
nav bar structure	27,251 → 28,160	+3.3%	76,162 → 77,943	+2.3%
card component tree	32,761 → 32,495	-0.8%	83,760 → 86,074	+2.8%
10 list items via Array.from	14,633 → 15,262	+4.3%	22,991 → 22,929	-0.3%

numbers.js — Interp: 🟢 1, 10 unch. · avg +2.7% · Bytecode: 🟢 4, 7 unch. · avg +5.5%

Benchmark	Interpreted	Δ	Bytecode	Δ
integer arithmetic	572,098 → 543,622	-5.0%	1,551,168 → 1,565,663	+0.9%
floating point arithmetic	581,760 → 613,350	+5.4%	1,704,377 → 1,705,066	+0.0%
number coercion	188,790 → 195,496	+3.6%	130,958 → 132,646	+1.3%
toFixed	103,966 → 108,805	+4.7%	212,084 → 219,217	+3.4%
toString	161,763 → 163,004	+0.8%	691,907 → 746,571	🟢 +7.9%
valueOf	232,794 → 244,520	+5.0%	922,329 → 1,067,235	🟢 +15.7%
toPrecision	148,448 → 145,984	-1.7%	402,664 → 419,367	+4.1%
Number.isNaN	305,747 → 319,096	+4.4%	167,637 → 175,702	+4.8%
Number.isFinite	300,438 → 305,897	+1.8%	164,320 → 172,490	+5.0%
Number.isInteger	290,481 → 311,563	🟢 +7.3%	173,849 → 187,915	🟢 +8.1%
Number.parseInt and parseFloat	246,387 → 255,356	+3.6%	144,373 → 157,403	🟢 +9.0%

objects.js — Interp: 7 unch. · avg -2.0% · Bytecode: 🟢 2, 5 unch. · avg +6.3%

Benchmark	Interpreted	Δ	Bytecode	Δ
create simple object	475,636 → 453,660	-4.6%	826,631 → 866,863	+4.9%
create nested object	218,984 → 223,459	+2.0%	361,623 → 383,939	+6.2%
create 50 objects via Array.from	9,424 → 9,487	+0.7%	8,358 → 8,292	-0.8%
property read	631,573 → 604,225	-4.3%	646,798 → 740,219	🟢 +14.4%
Object.keys	299,899 → 293,041	-2.3%	197,699 → 203,920	+3.1%
Object.entries	108,111 → 107,359	-0.7%	60,083 → 61,796	+2.9%
spread operator	195,104 → 186,178	-4.6%	175,599 → 198,561	🟢 +13.1%

promises.js — Interp: 🔴 2, 10 unch. · avg -3.9% · Bytecode: 🟢 2, 🔴 1, 9 unch. · avg +0.7%

Benchmark	Interpreted	Δ	Bytecode	Δ
Promise.resolve(value)	572,558 → 572,880	+0.1%	337,322 → 350,556	+3.9%
new Promise(resolve => resolve(value))	198,353 → 201,852	+1.8%	155,103 → 157,175	+1.3%
Promise.reject(reason)	554,217 → 538,808	-2.8%	334,101 → 340,277	+1.8%
resolve + then (1 handler)	174,111 → 170,718	-1.9%	147,480 → 158,510	🟢 +7.5%
resolve + then chain (3 deep)	68,657 → 62,233	🔴 -9.4%	66,475 → 62,079	-6.6%
resolve + then chain (10 deep)	21,947 → 19,314	🔴 -12.0%	23,314 → 21,612	🔴 -7.3%
reject + catch + then	99,537 → 95,216	-4.3%	86,362 → 95,299	🟢 +10.3%
resolve + finally + then	85,454 → 81,311	-4.8%	76,512 → 73,869	-3.5%
Promise.all (5 resolved)	32,761 → 31,164	-4.9%	26,910 → 25,804	-4.1%
Promise.race (5 resolved)	34,328 → 33,131	-3.5%	27,950 → 27,475	-1.7%
Promise.allSettled (5 mixed)	27,454 → 26,531	-3.4%	21,273 → 22,556	+6.0%
Promise.any (5 mixed)	31,475 → 30,793	-2.2%	25,947 → 25,997	+0.2%

strings.js — Interp: 🔴 1, 10 unch. · avg -3.4% · Bytecode: 🟢 5, 6 unch. · avg +6.6%

Benchmark	Interpreted	Δ	Bytecode	Δ
string concatenation	419,347 → 410,314	-2.2%	431,509 → 429,093	-0.6%
template literal	699,166 → 598,694	🔴 -14.4%	630,905 → 649,136	+2.9%
string repeat	419,083 → 417,903	-0.3%	925,354 → 1,021,021	🟢 +10.3%
split and join	138,048 → 131,933	-4.4%	312,660 → 315,872	+1.0%
indexOf and includes	171,761 → 171,040	-0.4%	607,863 → 648,516	+6.7%
toUpperCase and toLowerCase	261,106 → 251,618	-3.6%	629,490 → 689,471	🟢 +9.5%
slice and substring	164,329 → 157,630	-4.1%	660,254 → 732,949	🟢 +11.0%
trim operations	193,220 → 190,543	-1.4%	739,605 → 816,828	🟢 +10.4%
replace and replaceAll	212,068 → 203,020	-4.3%	671,880 → 705,192	+5.0%
startsWith and endsWith	139,914 → 139,580	-0.2%	503,897 → 555,097	🟢 +10.2%
padStart and padEnd	202,793 → 198,478	-2.1%	581,746 → 615,122	+5.7%

typed-arrays.js — Interp: 🟢 7, 15 unch. · avg +3.2% · Bytecode: 🟢 1, 21 unch. · avg +1.6%

Benchmark	Interpreted	Δ	Bytecode	Δ
new Int32Array(0)	334,573 → 315,273	-5.8%	125,768 → 128,188	+1.9%
new Int32Array(100)	306,590 → 301,525	-1.7%	121,655 → 123,750	+1.7%
new Int32Array(1000)	173,236 → 195,189	🟢 +12.7%	52,307 → 61,572	🟢 +17.7%
new Float64Array(100)	272,390 → 274,416	+0.7%	102,507 → 106,751	+4.1%
Int32Array.from([...])	183,138 → 197,505	🟢 +7.8%	153,812 → 151,964	-1.2%
Int32Array.of(1, 2, 3, 4, 5)	320,102 → 315,935	-1.3%	242,109 → 233,711	-3.5%
sequential write 100 elements	3,657 → 3,641	-0.4%	12,929 → 13,803	+6.8%
sequential read 100 elements	3,792 → 3,715	-2.0%	10,041 → 10,595	+5.5%
Float64Array write 100 elements	3,400 → 3,489	+2.6%	12,465 → 12,998	+4.3%
fill(42)	45,916 → 57,032	🟢 +24.2%	43,915 → 43,580	-0.8%
slice()	207,134 → 221,821	🟢 +7.1%	181,441 → 185,953	+2.5%
map(x => x * 2)	8,444 → 8,404	-0.5%	7,710 → 7,186	-6.8%
filter(x => x > 50)	8,592 → 8,510	-1.0%	7,895 → 7,558	-4.3%
reduce (sum)	8,054 → 8,016	-0.5%	6,656 → 6,566	-1.4%
sort()	177,016 → 206,604	🟢 +16.7%	148,590 → 154,319	+3.9%
indexOf()	461,971 → 503,561	🟢 +9.0%	348,035 → 359,390	+3.3%
reverse()	338,894 → 369,842	🟢 +9.1%	263,421 → 271,336	+3.0%
create view over existing buffer	413,022 → 403,430	-2.3%	133,493 → 137,789	+3.2%
subarray()	468,241 → 462,053	-1.3%	342,683 → 349,703	+2.0%
set() from array	604,335 → 629,768	+4.2%	255,037 → 237,765	-6.8%
for-of loop	5,214 → 4,983	-4.4%	18,335 → 17,787	-3.0%
spread into array	18,421 → 17,923	-2.7%	53,001 → 53,981	+1.8%

_{Measured on ubuntu-latest x64. Changes within ±7% are considered insignificant.}