Cytoscape.js GPU benchmarks

CPU	Intel(R) Core(TM) i9-9900K CPU @ 3.60GHz
Cores	8 physical / 16 logical · 3.6 GHz base · 5 GHz max
Memory	62.6 GiB
GPU	CoffeeLake-S GT2 [UHD Graphics 630] i915 Ellesmere [Radeon RX 470/480/570/570X/580/580X/590] (8.0 GiB) amdgpu
OS	Fedora Linux 43 (Workstation Edition) · 6.19.14-200.fc43.x86_64 · x64
Node	v24.18.0 (v8 13.6.233.17-node.50)
Commit	cf5727b4 perf(collection): _liveNodeRef reads the raw ref (round 62.6c)
Machine id	5cf3f79c

Failed jobs

style-bundle.mjs @ N=2000 (exit 1)

Speedup overview v3 p50 ÷ gpu p50, log scale · right of the 1× line = gpu faster

sweep: edges() @ 2k

290398×

sweep: nodes() @ 2k

171180×

core: edges() @ 2k

92914×

core: nodes() @ 2k

51238×

sweep: elements() @ 2k

29346×

core: elements() @ 2k

21280×

core: graph scratch() set + get (v3 restyles the graph) @ 2k

16571×

core: removeScratch() @ 2k

9678×

core: graph data() set + get (v3 restyles the graph) @ 2k

8610×

compare: same() @ 2k

7831×

trav: successors() @ 2k

781×

trav: predecessors() @ 2k

736×

mut-bulk: hide + show @ 2k

283×

flags: show + hide + visible @ 2k

224×

parent drag (shift subtree + bb settle) @ 2k

186×

degree: nodes.totalDegree() @ 2k

179×

degree: nodes.maxDegree() @ 2k

145×

compound: commonAncestors @ 2k

131×

degree: node.degree() @ 2k

127×

viewport: reset() @ 2k

121×

mut-bulk: positions(obj) @ 2k

112×

mut-bulk: shift @ 2k

100×

mut-bulk: positions(fn) @ 2k

97×

flags: selectify + unselectify + selectable @ 2k

95×

flags: panify + unpanify + pannable @ 2k

95×

flags: grabify + ungrabify + grabbable @ 2k

94×

flags: lock + unlock + locked @ 2k

92×

91×

core: options() @ 2k

86×

algo: closenessCentrality (one root) @ 500

77×

collection: same + anySame + contains (3 calls) @ 2k

72×

algo: closenessCentrality (one root) @ 2k

72×

degree: indegree + outdegree (2 calls) @ 2k

72×

reparent round-trip @ 2k

70×

viewport: viewport({ zoom, pan }) @ 2k

64×

algo: bfs (whole graph) @ 500

64×

algo: dfs (whole graph) @ 500

62×

dimensions: shift + silentShift (round-trip) @ 2k

60×

sweep: nodes(":selected") vs nodes({selected}) @ 2k

59×

algo: closenessCentrality (weighted, one root) @ 500

59×

algo: dfs (whole graph) @ 2k

59×

algo: dijkstra + pathTo @ 500

58×

57×

algo: dijkstra + pathTo @ 2k

56×

dimensions: positions( fn ) over 100 @ 2k

55×

algo: bfs (whole graph) @ 2k

55×

mut-bulk: lock + unlock @ 2k

52×

child drag (move + parent re-derive) @ 2k

51×

collection: component() @ 2k

42×

algo: degreeCentralityNormalized @ 500

40×

traverse: components() @ 2k

40×

algo: degreeCentralityNormalized @ 2k

39×

algo: tarjan SCC @ 500

38×

algo: tarjan SCC @ 2k

36×

sweep: selected nodes ($("node:selected") vs filter({group, selected})) @ 2k

36×

collection: silentPositions( fn ) over 100 @ 2k

35×

35×

degree: minDegree + maxDegree (2 calls) @ 2k

35×

algo: degreeCentrality (one root) @ 500

34×

bounds: getFitViewport() (compute without committing) @ 2k

33×

query: structural, parents (:parent vs { parent: true }) @ 2k

33×

bounds: cy.fit() (whole-graph scan) @ 2k

33×

query: structural, orphans (:orphan vs { child: false }) @ 2k

33×

algo: degreeCentrality (one root) @ 2k

30×

emit: position write, no listeners @ 2k

29×

collection: difference @ 2k

28×

scn: select-all + fit (listeners on) @ 2k

28×

algo: hopcroft-tarjan biconnected @ 500

28×

core: id lookup ($("#id") vs $id()) @ 2k

28×

mut: position set @ 2k

27×

collection: intersection @ 2k

27×

algo: kruskal @ 2k

27×

26×

algo: kruskal @ 500

26×

algo: hopcroft-tarjan biconnected @ 2k

25×

set: intersection() @ 2k

24×

set: difference() @ 2k

24×

trav: band roots() @ 2k

23×

degree: maxIndegree + minOutdegree (2 calls) @ 2k

22×

degree: minIndegree + maxOutdegree (2 calls) @ 2k

21×

data: bulk write, numeric column @ 2k

21×

algo: bellmanFord @ 500

21×

data: bulk write, object column (plain-array fallback) @ 2k

20×

viewport: center( eles ) @ 2k

19×

19×

algo: bellmanFord @ 2k

19×

core: add + remove @ 2k

17×

dimensions: silentPosition @ 2k

17×

data: bulk write, one new dictionary entry per pass @ 2k

17×

data: bulk write, dictionary string column (4 values) @ 2k

17×

scn: drag (select band of 100 + 8-step shift, position listener) @ 2k

16×

core: forceRender() (headless no-op) @ 2k

16×

query: membership ([bucket in ...] vs { in }) @ 2k

15×

dimensions: padding + paddedWidth + paddedHeight (3 calls) @ 2k

15×

flags: select + unselect + selected (band) @ 2k

15×

flags: active/activate/unactivate (3 calls) @ 2k

15×

traversal: roots + leaves (2 calls) @ 2k

15×

viewport: minZoom/maxZoom get (2 calls) @ 2k

15×

mut-bulk: data set @ 2k

15×

style: renderedStyle (one prop) @ 2k

14×

core: selected ($(":selected") vs filter({selected})) @ 2k

14×

sweep: selected ($(":selected") vs filter({selected})) @ 2k

14×

query: structural, childless (:childless vs { parent: false }) @ 2k

13×

collection: absoluteComplement @ 2k

13×

query: data equality ([bucket = 3] vs { data: { bucket: 3 } }) @ 2k

13×

query: data comparison ([weight > 3] vs { gt }) @ 2k

12×

12×

collection: toArray() @ 2k

12×

query: two keys AND-ed @ 2k

12×

algo: betweennessCentrality (unweighted) @ 2k

12×

dimensions: relativePosition get @ 2k

12×

algo: betweennessCentrality (unweighted) @ 500

11×

11×

9.5×

collection: removeScratch() @ 2k

9.0×

emit: delegated (v3 selector vs v4 predicate) @ 2k

8.9×

sweep: all nodes ($("node") vs filter({group})) @ 2k

8.9×

algo: betweennessCentrality (weighted) @ 500

8.9×

core: removeAllListeners() @ 2k

8.8×

mut-bulk: select + unselect @ 2k

8.7×

algo: kMedoids @ 500

8.6×

core: batchData() @ 2k

8.6×

emit: 1 core listener @ 2k

8.5×

emit: 1 element (ref-qualified) listener @ 2k

8.3×

algo: fuzzyCMeans @ 500

8.1×

scn: explore (expand 2-hop + select + fit) @ 2k

7.9×

data: removeData (whole collection) @ 2k

7.7×

element data: scratch get + set @ 2k

7.3×

core: collection() @ 2k

7.2×

scn: edit (remove + re-add 256 nodes + cascade, add/remove listeners) @ 2k

7.2×

core: collection() @ 2k

7.0×

emit: 10 core listeners @ 2k

6.9×

core: batch( fn ) (empty) @ 2k

6.8×

query: predicate function (both sides) @ 2k

6.7×

compound: orphans + nonorphans (2 calls) @ 2k

6.6×

collection: union @ 2k

6.3×

style: whole-object style() @ 2k

6.2×

bounds: whole-graph boundingBox() @ 2k

6.1×

mut-bulk: remove + re-add (256 nodes + cascade) @ 2k

6.0×

load: init from the definition form (construct + dispose) @ 2k

5.9×

set: union() @ 2k

5.9×

core: startBatch + endBatch + batching @ 2k

5.8×

core: nodes ($("node") vs filter({group})) @ 2k

5.6×

algo: hierholzer (directed) @ 500

5.6×

trav: outgoers() @ 2k

5.1×

trav: incomers() @ 2k

5.1×

traversal: allAreNeighbors @ 2k

5.1×

trav: band neighborhood() @ 2k

5.0×

core: animated() @ 2k

4.9×

traverse: incomers() @ 2k

4.9×

traverse: outgoers() @ 2k

4.8×

compare: contains() @ 2k

4.8×

viewport: zoom() get @ 2k

4.7×

compound: children + descendants + siblings (3 calls) @ 2k

4.6×

collection: length + empty + nonempty (3 calls) @ 2k

4.5×

mut: data set @ 2k

4.1×

mut: select + unselect @ 2k

4.0×

trav: neighborhood() @ 2k

3.9×

algo: kMeans (fixed centroids) @ 500

3.8×

collection: symmetricDifference @ 2k

3.7×

compound: parent + parents + ancestors (3 calls) @ 2k

3.6×

collection: layoutDimensions() @ 2k

3.6×

algo: aStar @ 500

3.6×

collection: cy() @ 2k

3.5×

traverse: neighborhood() @ 2k

3.5×

traversal: sources + targets (2 calls) @ 2k

3.4×

traversal: parallelEdges + codirectedEdges (2 calls) @ 2k

3.3×

algo: aStar @ 2k

3.3×

traversal: connectedNodes @ 2k

3.3×

anim: start + stop, 512-node collection @ 2k

3.0×

dimensions: rendered dims (4 calls) @ 2k

2.9×

traversal: openNeighborhood + closedNeighborhood (2 calls) @ 2k

2.9×

collection: diff @ 2k

2.9×

collection: animated() @ 2k

2.9×

flags: takesUpSpace + interactive + transparent (3 calls) @ 2k

2.9×

viewport: width() + height() + size() (3 calls) @ 2k

2.9×

trav: band connectedNodes() @ 2k

2.8×

algo: hierholzer (directed) @ 2k

2.8×

core: window() @ 2k

2.8×

anim: viewport animate + stop @ 2k

2.8×

style: numericStyle @ 2k

2.8×

trav: closedNeighborhood() @ 2k

2.7×

core: resize() / invalidateSize() @ 2k

2.7×

collection: delayAnimation() @ 2k

2.7×

load: cy.json() @ 2k

2.7×

core: container() @ 2k

2.7×

collection: byGroup @ 2k

2.6×

dimensions: renderedPosition get @ 2k

2.5×

trav: connectedEdges() @ 2k

2.5×

dimensions: width + height + outerWidth + outerHeight (4 calls) @ 2k

2.5×

iter: map() @ 2k

2.4×

core: onRender + offRender @ 2k

2.4×

traverse: connectedEdges() @ 2k

2.4×

traversal: isLoop + isSimple (2 calls) @ 2k

2.3×

algo: hierarchicalClustering (threshold) @ 500

2.3×

data: read the whole object @ 2k

2.3×

core: isReady() @ 2k

2.2×

core: destroyed() @ 2k

2.2×

emit: on() + off() round-trip @ 2k

2.2×

collection: is (none match: the full walk) @ 2k

2.2×

flags: effectiveOpacity @ 2k

2.1×

collection: slice() @ 2k

2.1×

style: read one prop (width) @ 2k

2.1×

collection: min + max ( fn ) (2 calls) @ 2k

2.0×

flags: grabbed + hidden + inactive (3 calls) @ 2k

2.0×

core: selectionType get @ 2k

1.99×

element data: json() one element @ 2k

1.98×

json: element @ 2k

1.98×

viewport: pan( p ) set @ 2k

1.95×

core: zoom set @ 2k

1.93×

dimensions: renderedBoundingBox @ 2k

1.91×

style: read one prop (color) @ 2k

1.90×

algo: pageRank (20 iters) @ 500

1.89×

collection: indexOfId() @ 2k

1.87×

element data: jsons() over 100 @ 2k

1.87×

core: gating setters round-trip (autolock/autoungrabify/autounselectify) @ 2k

1.85×

scn: refresh (bulk data write + mapped labels + filter + fit, data listener) @ 2k

1.84×

collection: eq/first/last (3 calls) @ 2k

1.83×

viewport: zoom( n ) set @ 2k

1.80×

collection: element accessors (id/group/isNode/isEdge, 4 calls) @ 2k

1.79×

algo: floydWarshall @ 500

1.78×

algo: affinityPropagation @ 500

1.77×

bounds: one node boundingBox() @ 2k

1.75×

viewport: panBy() @ 2k

1.73×

sweep: filter(fn) @ 2k

1.70×

algo: closenessCentralityNormalized @ 500

1.68×

iter: filter(fn) @ 2k

1.67×

viewport: renderedExtent() @ 2k

1.65×

viewport: extent() @ 2k

1.58×

anim: delay() (no channels touched) @ 2k

1.58×

compound: isParent/isChild/isChildless/isOrphan (4 calls) @ 2k

1.57×

collection: sort( fn ) @ 2k

1.57×

style: first apply on add (256-node band) @ 2k

1.57×

collection: some + every (2 calls) @ 2k

1.57×

collection: allAre (all match: the full walk) @ 2k

1.52×

position: get @ 2k

1.49×

trav: band sources() @ 2k

1.48×

trav: band connectedEdges() @ 2k

1.45×

collection: indexOf @ 2k

1.41×

collection: reduce( fn ) @ 2k

1.38×

anim: start + stop, one element @ 2k

1.37×

load: add a 256-node band (add + remove) @ 2k

1.36×

data: read one numeric key @ 2k

1.36×

core: mutableElements() (v4 materializes; v3 is O(1)) @ 2k

1.36×

core: pan() get (x32) @ 2k

1.35×

data: read one string key (dictionary decode) @ 2k

1.33×

core: hasElementWithId + hasCompoundNodes (2 calls) @ 2k

1.27×

traversal: source() + target() (2 calls) @ 2k

1.27×

algo: markovClustering @ 500

1.25×

core: instanceString + headless + styleEnabled (3 calls) @ 2k

1.25×

traversal: edgesWith @ 2k

1.21×

core: filter(fn) @ 2k

1.21×

style: sheet swap (compile + applyAll) @ 2k

1.20×

traversal: edgesTo @ 2k

1.17×

core: gating getters (5 calls) @ 2k

1.16×

core: getElementById() @ 2k

1.12×

collection: removed() + inside() (2 calls) @ 2k

1.11×

core: extent() @ 2k

1.09×

data: get @ 2k

1.09×

viewport: pan() get @ 2k

1.08×

iter: forEach() @ 2k

1.05×

algo: kargerStein @ 500

1.03×

1×10×100×1000×10000×100000×1000000×

Scaling speedup vs graph size

algorithms

group	× @ N=500	× @ N=2k
algo: bfs (whole graph)	64×	55×
algo: dfs (whole graph)	62×	59×
algo: dijkstra + pathTo	58×	56×
algo: aStar	3.6×	3.3×
algo: bellmanFord	21×	19×
algo: kruskal	26×	27×
algo: tarjan SCC	38×	36×
algo: hopcroft-tarjan biconnected	28×	25×
algo: hierholzer (directed)	5.6×	2.8×
algo: betweennessCentrality (unweighted)	11×	12×
algo: degreeCentralityNormalized	40×	39×
algo: closenessCentrality (one root)	77×	72×
algo: degreeCentrality (one root)	34×	30×
algo: pageRank (20 iters)	1.89×	–
algo: closenessCentrality (weighted, one root)	59×	–
algo: floydWarshall	1.78×	–
algo: closenessCentralityNormalized	1.68×	–
algo: markovClustering	1.25×	–
algo: hierarchicalClustering (threshold)	2.3×	–
algo: kMeans (fixed centroids)	3.8×	–
algo: kMedoids	8.6×	–
algo: fuzzyCMeans	8.1×	–
algo: affinityPropagation	1.77×	–
algo: kargerStein	1.03×	–
algo: betweennessCentrality (weighted)	8.9×	–

core+collection N=2,000 · 57.8 sv4 (gpu) v3

core: getElementById()

1.12×

core: id lookup ($("#id") vs $id())

28×

core: nodes()

51238×

core: edges()

92914×

core: elements()

21280×

core: nodes ($("node") vs filter({group}))

5.6×

core: selected ($(":selected") vs filter({selected}))

14×

core: filter(fn)

1.21×

core: collection()

7.0×

core: pan() get (x32)

1.35×

core: extent()

1.09×

core: add + remove

17×

core: zoom set

1.93×

traverse: neighborhood()

3.5×

traverse: connectedEdges()

2.4×

traverse: outgoers()

4.8×

traverse: incomers()

4.9×

traverse: components()

40×

degree: node.degree()

127×

degree: nodes.totalDegree()

179×

degree: nodes.maxDegree()

145×

data: get

1.09×

position: get

1.49×

json: element

1.98×

set: union()

5.9×

set: intersection()

24×

set: difference()

24×

iter: map()

2.4×

iter: forEach()

1.05×

iter: filter(fn)

1.67×

compare: same()

7831×

compare: contains()

4.8×

mut: data set

4.1×

mut: position set

27×

mut: select + unselect

4.0×

1 ns10 ns100 ns1 µs10 µs100 µs1 ms10 ms100 ms

All stats (table view)

group	bench	p50	avg	p99	min	samples
core: getElementById()	v3	96.2 ns	99.9 ns	158 ns	92.8 ns	1714
core: getElementById()	gpu	86.2 ns	91.7 ns	164 ns	84.4 ns	1867
core: id lookup ($("#id") vs $id())	v3	2.19 µs	2.45 µs	3.83 µs	1.94 µs	63
core: id lookup ($("#id") vs $id())	gpu	79.4 ns	85.8 ns	154 ns	77.6 ns	1991
core: nodes()	v3	421 µs	449 µs	820 µs	409 µs	1558
core: nodes()	gpu	8.21 ns	9.18 ns	31.2 ns	7.88 ns	18483
core: edges()	v3	733 µs	776 µs	1.45 ms	716 µs	899
core: edges()	gpu	7.89 ns	8.65 ns	16.5 ns	7.72 ns	19589
core: elements()	v3	147 µs	157 µs	275 µs	144 µs	4442
core: elements()	gpu	6.93 ns	7.40 ns	17.5 ns	6.77 ns	22941
core: nodes ($("node") vs filter({group}))	v3	553 µs	589 µs	1.21 ms	536 µs	1184
core: nodes ($("node") vs filter({group}))	gpu	98.3 µs	107 µs	259 µs	65.9 µs	6368
core: selected ($(":selected") vs filter({selected}))	v3	347 µs	368 µs	729 µs	340 µs	1860
core: selected ($(":selected") vs filter({selected}))	gpu	25.3 µs	26.2 µs	27.1 µs	25.1 µs	12
core: filter(fn)	v3	416 µs	437 µs	946 µs	411 µs	1597
core: filter(fn)	gpu	344 µs	373 µs	751 µs	340 µs	1846
core: collection()	v3	477 ns	562 ns	1.59 µs	453 ns	292
core: collection()	gpu	68.2 ns	71.7 ns	123 ns	67.1 ns	2276
core: pan() get (x32)	v3	45.6 ns	47.9 ns	83.0 ns	43.4 ns	3552
core: pan() get (x32)	gpu	33.8 ns	36.2 ns	53.1 ns	32.9 ns	4724
core: extent()	v3	22.7 ns	40.3 ns	144 ns	22.0 ns	4123
core: extent()	gpu	20.7 ns	27.7 ns	92.5 ns	19.3 ns	6029
core: add + remove	v3	120 µs	127 µs	236 µs	110 µs	5428
core: add + remove	gpu	6.96 µs	8.12 µs	24.9 µs	6.54 µs	78345
core: zoom set	v3	1.28 µs	1.29 µs	1.45 µs	1.21 µs	125
core: zoom set	gpu	662 ns	685 ns	1.07 µs	641 ns	238
traverse: neighborhood()	v3	3.61 µs	3.92 µs	4.86 µs	3.50 µs	40
traverse: neighborhood()	gpu	1.03 µs	1.18 µs	2.36 µs	998 ns	136
traverse: connectedEdges()	v3	1.05 µs	1.16 µs	2.01 µs	1.02 µs	143
traverse: connectedEdges()	gpu	442 ns	494 ns	1.42 µs	431 ns	339
traverse: outgoers()	v3	2.38 µs	2.55 µs	3.62 µs	2.27 µs	63
traverse: outgoers()	gpu	495 ns	604 ns	2.07 µs	483 ns	268
traverse: incomers()	v3	2.42 µs	2.66 µs	3.87 µs	2.33 µs	59
traverse: incomers()	gpu	497 ns	611 ns	2.11 µs	485 ns	264
traverse: components()	v3	79.7 ms	81.7 ms	83.9 ms	76.3 ms	12
traverse: components()	gpu	2.02 ms	2.12 ms	3.25 ms	1.94 ms	315
degree: node.degree()	v3	3.63 µs	3.83 µs	5.09 µs	3.30 µs	41
degree: node.degree()	gpu	28.7 ns	30.4 ns	58.1 ns	27.4 ns	5626
degree: nodes.totalDegree()	v3	7.81 ms	8.70 ms	13.0 ms	7.38 ms	76
degree: nodes.totalDegree()	gpu	43.7 µs	45.1 µs	85.8 µs	42.2 µs	15268
degree: nodes.maxDegree()	v3	7.47 ms	8.31 ms	11.8 ms	7.09 ms	80
degree: nodes.maxDegree()	gpu	51.6 µs	53.3 µs	91.1 µs	49.1 µs	13078
data: get	v3	30.1 ns	31.4 ns	53.2 ns	28.8 ns	5455
data: get	gpu	27.6 ns	28.4 ns	39.3 ns	26.4 ns	6015
position: get	v3	31.0 ns	32.7 ns	67.4 ns	30.4 ns	5230
position: get	gpu	20.9 ns	24.9 ns	54.3 ns	20.2 ns	6723
json: element	v3	303 ns	391 ns	3.59 µs	295 ns	419
json: element	gpu	153 ns	159 ns	240 ns	149 ns	1021
set: union()	v3	102 µs	112 µs	167 µs	94.2 µs	6020
set: union()	gpu	17.2 µs	17.8 µs	18.4 µs	16.9 µs	12
set: intersection()	v3	92.3 µs	105 µs	156 µs	88.0 µs	6566
set: intersection()	gpu	3.84 µs	3.88 µs	4.16 µs	3.74 µs	38
set: difference()	v3	93.3 µs	107 µs	179 µs	88.4 µs	6418
set: difference()	gpu	3.92 µs	4.01 µs	4.59 µs	3.84 µs	37
iter: map()	v3	27.1 µs	29.5 µs	74.9 µs	26.0 µs	23242
iter: map()	gpu	11.1 µs	11.1 µs	11.2 µs	11.0 µs	11
iter: forEach()	v3	3.48 µs	3.75 µs	4.15 µs	3.44 µs	36
iter: forEach()	gpu	3.31 µs	3.35 µs	3.54 µs	3.27 µs	45
iter: filter(fn)	v3	245 µs	255 µs	385 µs	243 µs	2740
iter: filter(fn)	gpu	147 µs	154 µs	256 µs	142 µs	4539
compare: same()	v3	117 µs	120 µs	173 µs	112 µs	5823
compare: same()	gpu	14.9 ns	15.6 ns	26.0 ns	14.1 ns	10927
compare: contains()	v3	6.01 µs	6.00 µs	6.15 µs	5.88 µs	25
compare: contains()	gpu	1.25 µs	1.26 µs	1.36 µs	1.23 µs	129
mut: data set	v3	445 ns	462 ns	734 ns	432 ns	361
mut: data set	gpu	108 ns	114 ns	169 ns	105 ns	1484
mut: position set	v3	1.30 µs	1.32 µs	1.49 µs	1.27 µs	123
mut: position set	gpu	47.6 ns	50.1 ns	98.7 ns	46.3 ns	3306
mut: select + unselect	v3	2.76 µs	2.91 µs	3.44 µs	2.59 µs	53
mut: select + unselect	gpu	686 ns	743 ns	1.21 µs	628 ns	680626

materializers N=2,000 · 16.6 sv4 (gpu) v3

sweep: nodes()

171180×

sweep: edges()

290398×

sweep: elements()

29346×

sweep: all nodes ($("node") vs filter({group}))

8.9×

sweep: selected ($(":selected") vs filter({selected}))

14×

sweep: selected nodes ($("node:selected") vs filter({group, selected}))

36×

sweep: nodes(":selected") vs nodes({selected})

59×

sweep: filter(fn)

1.70×

1 ns10 ns100 ns1 µs10 µs100 µs1 ms

All stats (table view)

group	bench	p50	avg	p99	min	samples
sweep: nodes()	v3	412 µs	424 µs	520 µs	400 µs	1654
sweep: nodes()	gpu	2.41 ns	2.49 ns	4.76 ns	2.30 ns	66454
sweep: edges()	v3	729 µs	743 µs	972 µs	708 µs	941
sweep: edges()	gpu	2.51 ns	2.61 ns	3.14 ns	2.51 ns	63759
sweep: elements()	v3	143 µs	147 µs	207 µs	139 µs	4745
sweep: elements()	gpu	4.86 ns	5.00 ns	11.2 ns	4.54 ns	33635
sweep: all nodes ($("node") vs filter({group}))	v3	544 µs	573 µs	862 µs	536 µs	1219
sweep: all nodes ($("node") vs filter({group}))	gpu	61.2 µs	73.6 µs	183 µs	53.0 µs	9409
sweep: selected ($(":selected") vs filter({selected}))	v3	349 µs	361 µs	578 µs	342 µs	1937
sweep: selected ($(":selected") vs filter({selected}))	gpu	25.8 µs	26.2 µs	26.9 µs	25.4 µs	12
sweep: selected nodes ($("node:selected") vs filter({group, selected}))	v3	322 µs	334 µs	534 µs	314 µs	2093
sweep: selected nodes ($("node:selected") vs filter({group, selected}))	gpu	9.01 µs	9.04 µs	9.32 µs	8.89 µs	15
sweep: nodes(":selected") vs nodes({selected})	v3	533 µs	549 µs	903 µs	517 µs	1274
sweep: nodes(":selected") vs nodes({selected})	gpu	8.97 µs	8.99 µs	9.21 µs	8.87 µs	15
sweep: filter(fn)	v3	412 µs	421 µs	498 µs	400 µs	1659
sweep: filter(fn)	gpu	242 µs	258 µs	570 µs	236 µs	2696

mutators N=2,000 · 18.0 sv4 (gpu) v3

mut-bulk: select + unselect

8.7×

mut-bulk: hide + show

283×

mut-bulk: lock + unlock

52×

mut-bulk: positions(obj)

112×

mut-bulk: positions(fn)

97×

mut-bulk: shift

100×

mut-bulk: data set

15×

mut-bulk: remove + re-add (256 nodes + cascade)

6.0×

10 µs100 µs1 ms10 ms100 ms

All stats (table view)

group	bench	p50	avg	p99	min	samples
mut-bulk: select + unselect	v3	2.21 ms	2.24 ms	3.48 ms	2.02 ms	295
mut-bulk: select + unselect	gpu	254 µs	261 µs	387 µs	246 µs	2685
mut-bulk: hide + show	v3	60.4 ms	66.7 ms	73.8 ms	58.8 ms	12
mut-bulk: hide + show	gpu	214 µs	224 µs	738 µs	203 µs	3089
mut-bulk: lock + unlock	v3	1.99 ms	2.02 ms	2.39 ms	1.95 ms	330
mut-bulk: lock + unlock	gpu	38.5 µs	38.9 µs	39.2 µs	38.2 µs	12
mut-bulk: positions(obj)	v3	2.28 ms	2.33 ms	4.01 ms	2.19 ms	293
mut-bulk: positions(obj)	gpu	20.4 µs	21.9 µs	47.2 µs	19.3 µs	31436
mut-bulk: positions(fn)	v3	3.37 ms	3.56 ms	4.98 ms	3.24 ms	190
mut-bulk: positions(fn)	gpu	34.8 µs	38.0 µs	101 µs	33.0 µs	18168
mut-bulk: shift	v3	3.41 ms	3.58 ms	4.88 ms	3.27 ms	189
mut-bulk: shift	gpu	34.3 µs	36.5 µs	66.8 µs	33.0 µs	19002
mut-bulk: data set	v3	919 µs	949 µs	1.58 ms	875 µs	722
mut-bulk: data set	gpu	62.4 µs	65.2 µs	136 µs	59.9 µs	10524
mut-bulk: remove + re-add (256 nodes + cascade)	v3	41.2 ms	43.0 ms	47.6 ms	40.2 ms	12
mut-bulk: remove + re-add (256 nodes + cascade)	gpu	6.89 ms	7.01 ms	8.33 ms	6.17 ms	93

scenarios N=2,000 · 12.7 sv4 (gpu) v3

scn: explore (expand 2-hop + select + fit)

7.9×

scn: select-all + fit (listeners on)

28×

scn: drag (select band of 100 + 8-step shift, position listener)

16×

scn: edit (remove + re-add 256 nodes + cascade, add/remove listeners)

7.2×

scn: refresh (bulk data write + mapped labels + filter + fit, data listener)

1.84×

10 µs100 µs1 ms10 ms100 ms

All stats (table view)

group	bench	p50	avg	p99	min	samples
scn: explore (expand 2-hop + select + fit)	v3	297 µs	349 µs	1.01 ms	279 µs	1988
scn: explore (expand 2-hop + select + fit)	gpu	37.6 µs	40.8 µs	82.2 µs	34.5 µs	16286
scn: select-all + fit (listeners on)	v3	25.3 ms	24.9 ms	26.6 ms	23.3 ms	23
scn: select-all + fit (listeners on)	gpu	910 µs	955 µs	1.82 ms	843 µs	700
scn: drag (select band of 100 + 8-step shift, position listener)	v3	1.64 ms	1.72 ms	2.47 ms	1.58 ms	389
scn: drag (select band of 100 + 8-step shift, position listener)	gpu	99.9 µs	108 µs	323 µs	96.4 µs	6297
scn: edit (remove + re-add 256 nodes + cascade, add/remove listeners)	v3	50.9 ms	50.6 ms	52.1 ms	46.7 ms	10
scn: edit (remove + re-add 256 nodes + cascade, add/remove listeners)	gpu	7.08 ms	7.32 ms	8.98 ms	6.44 ms	90
scn: refresh (bulk data write + mapped labels + filter + fit, data listener)	v3	7.72 ms	8.28 ms	13.4 ms	7.24 ms	79
scn: refresh (bulk data write + mapped labels + filter + fit, data listener)	gpu	4.20 ms	4.33 ms	6.93 ms	3.75 ms	149

traversal N=2,000 · 40.6 sv4 (gpu) v3

trav: neighborhood()

3.9×

trav: closedNeighborhood()

2.7×

trav: connectedEdges()

2.5×

trav: outgoers()

5.1×

trav: incomers()

5.1×

trav: successors()

781×

trav: predecessors()

736×

trav: band neighborhood()

5.0×

trav: band connectedEdges()

1.45×

trav: band roots()

23×

trav: band sources()

1.48×

trav: band connectedNodes()

2.8×

100 ns1 µs10 µs100 µs1 ms10 ms100 ms1 s

All stats (table view)

group	bench	p50	avg	p99	min	samples
trav: neighborhood()	v3	3.91 µs	4.10 µs	5.43 µs	3.43 µs	37
trav: neighborhood()	gpu	1.01 µs	1.08 µs	2.10 µs	853 ns	154
trav: closedNeighborhood()	v3	5.93 µs	5.74 µs	6.15 µs	5.23 µs	26
trav: closedNeighborhood()	gpu	2.16 µs	2.34 µs	2.97 µs	2.09 µs	69
trav: connectedEdges()	v3	978 ns	1.06 µs	1.79 µs	954 ns	158
trav: connectedEdges()	gpu	397 ns	414 ns	845 ns	388 ns	408
trav: outgoers()	v3	2.32 µs	2.49 µs	3.09 µs	2.25 µs	65
trav: outgoers()	gpu	452 ns	534 ns	2.06 µs	442 ns	310
trav: incomers()	v3	2.28 µs	2.45 µs	3.02 µs	2.21 µs	65
trav: incomers()	gpu	448 ns	532 ns	2.01 µs	442 ns	313
trav: successors()	v3	442 ms	446 ms	462 ms	433 ms	12
trav: successors()	gpu	565 µs	595 µs	1.90 ms	547 µs	1145
trav: predecessors()	v3	437 ms	440 ms	458 ms	431 ms	12
trav: predecessors()	gpu	593 µs	619 µs	1.83 ms	577 µs	1098
trav: band neighborhood()	v3	190 µs	204 µs	399 µs	183 µs	3415
trav: band neighborhood()	gpu	38.1 µs	38.7 µs	41.0 µs	37.3 µs	12
trav: band connectedEdges()	v3	24.7 µs	25.4 µs	26.4 µs	24.3 µs	12
trav: band connectedEdges()	gpu	17.1 µs	17.9 µs	19.6 µs	16.7 µs	12
trav: band roots()	v3	111 µs	120 µs	188 µs	106 µs	5773
trav: band roots()	gpu	4.89 µs	4.89 µs	5.07 µs	4.75 µs	31
trav: band sources()	v3	6.58 µs	6.55 µs	6.72 µs	6.30 µs	22
trav: band sources()	gpu	4.43 µs	4.52 µs	4.90 µs	4.35 µs	34
trav: band connectedNodes()	v3	14.1 µs	14.4 µs	14.8 µs	13.8 µs	12
trav: band connectedNodes()	gpu	4.98 µs	5.02 µs	5.17 µs	4.90 µs	30

algorithms N=2,000 · 1.5 minv4 (gpu) v3

algo: bfs (whole graph)

55×

algo: dfs (whole graph)

59×

algo: dijkstra + pathTo

56×

algo: aStar

3.3×

algo: bellmanFord

19×

algo: kruskal

27×

algo: tarjan SCC

36×

algo: hopcroft-tarjan biconnected

25×

algo: hierholzer (directed)

2.8×

algo: betweennessCentrality (unweighted)

12×

algo: degreeCentralityNormalized

39×

algo: closenessCentrality (one root)

72×

algo: degreeCentrality (one root)

30×

100 ns1 µs10 µs100 µs1 ms10 ms100 ms1 s10 s

All stats (table view)

group	bench	p50	avg	p99	min	samples
algo: bfs (whole graph)	v3	50.3 ms	52.9 ms	60.0 ms	48.7 ms	9
algo: bfs (whole graph)	gpu	915 µs	927 µs	1.44 ms	851 µs	747
algo: dfs (whole graph)	v3	48.5 ms	48.4 ms	48.9 ms	47.9 ms	11
algo: dfs (whole graph)	gpu	825 µs	844 µs	1.14 ms	808 µs	813
algo: dijkstra + pathTo	v3	51.4 ms	52.2 ms	54.1 ms	50.9 ms	10
algo: dijkstra + pathTo	gpu	925 µs	941 µs	1.12 ms	906 µs	729
algo: aStar	v3	1.65 ms	1.69 ms	2.25 ms	1.62 ms	410
algo: aStar	gpu	505 µs	530 µs	886 µs	488 µs	1316
algo: bellmanFord	v3	226 ms	226 ms	246 ms	206 ms	12
algo: bellmanFord	gpu	12.2 ms	12.2 ms	14.0 ms	11.8 ms	53
algo: kruskal	v3	13.6 ms	14.9 ms	19.3 ms	12.6 ms	42
algo: kruskal	gpu	506 µs	520 µs	826 µs	464 µs	1332
algo: tarjan SCC	v3	23.3 ms	24.0 ms	27.7 ms	21.5 ms	25
algo: tarjan SCC	gpu	652 µs	818 µs	5.29 ms	617 µs	826
algo: hopcroft-tarjan biconnected	v3	85.9 ms	87.0 ms	91.5 ms	81.8 ms	12
algo: hopcroft-tarjan biconnected	gpu	3.44 ms	3.54 ms	4.65 ms	3.38 ms	190
algo: hierholzer (directed)	v3	22.4 ms	23.5 ms	26.6 ms	21.2 ms	26
algo: hierholzer (directed)	gpu	8.07 ms	8.25 ms	11.0 ms	7.74 ms	75
algo: betweennessCentrality (unweighted)	v3	3.98 s	3.97 s	4.02 s	3.91 s	12
algo: betweennessCentrality (unweighted)	gpu	344 ms	363 ms	433 ms	334 ms	12
algo: degreeCentralityNormalized	v3	15.2 ms	16.0 ms	19.5 ms	14.0 ms	40
algo: degreeCentralityNormalized	gpu	392 µs	405 µs	661 µs	386 µs	1725
algo: closenessCentrality (one root)	v3	58.9 ms	61.3 ms	69.4 ms	57.4 ms	12
algo: closenessCentrality (one root)	gpu	816 µs	842 µs	1.18 ms	803 µs	827
algo: degreeCentrality (one root)	v3	7.33 µs	7.25 µs	8.21 µs	6.55 µs	20
algo: degreeCentrality (one root)	gpu	246 ns	252 ns	332 ns	240 ns	671

algorithms N=500 · 6.9 minv4 (gpu) v3

algo: bfs (whole graph)

64×

algo: dfs (whole graph)

62×

algo: dijkstra + pathTo

58×

algo: aStar

3.6×

algo: bellmanFord

21×

algo: kruskal

26×

algo: tarjan SCC

38×

algo: hopcroft-tarjan biconnected

28×

algo: hierholzer (directed)

5.6×

algo: betweennessCentrality (unweighted)

11×

algo: degreeCentralityNormalized

40×

algo: closenessCentrality (one root)

77×

algo: degreeCentrality (one root)

34×

algo: pageRank (20 iters)

1.89×

algo: closenessCentrality (weighted, one root)

59×

algo: floydWarshall

1.78×

algo: closenessCentralityNormalized

1.68×

algo: markovClustering

1.25×

algo: hierarchicalClustering (threshold)

2.3×

algo: kMeans (fixed centroids)

3.8×

algo: kMedoids

8.6×

algo: fuzzyCMeans

8.1×

algo: affinityPropagation

1.77×

algo: kargerStein

1.03×

algo: betweennessCentrality (weighted)

8.9×

100 ns1 µs10 µs100 µs1 ms10 ms100 ms1 s10 s

All stats (table view)

group	bench	p50	avg	p99	min	samples
algo: bfs (whole graph)	v3	13.5 ms	14.0 ms	19.3 ms	11.6 ms	45
algo: bfs (whole graph)	gpu	212 µs	218 µs	358 µs	195 µs	3154
algo: dfs (whole graph)	v3	12.1 ms	13.1 ms	16.5 ms	11.6 ms	50
algo: dfs (whole graph)	gpu	194 µs	206 µs	355 µs	190 µs	3379
algo: dijkstra + pathTo	v3	13.3 ms	14.1 ms	17.2 ms	12.4 ms	46
algo: dijkstra + pathTo	gpu	230 µs	236 µs	336 µs	221 µs	2955
algo: aStar	v3	439 µs	448 µs	589 µs	430 µs	1556
algo: aStar	gpu	123 µs	128 µs	197 µs	118 µs	5450
algo: bellmanFord	v3	16.9 ms	17.0 ms	18.1 ms	16.5 ms	37
algo: bellmanFord	gpu	808 µs	838 µs	1.93 ms	786 µs	830
algo: kruskal	v3	3.42 ms	3.71 ms	7.46 ms	3.31 ms	185
algo: kruskal	gpu	131 µs	140 µs	236 µs	125 µs	4982
algo: tarjan SCC	v3	5.71 ms	6.09 ms	9.18 ms	5.52 ms	110
algo: tarjan SCC	gpu	151 µs	164 µs	303 µs	144 µs	4254
algo: hopcroft-tarjan biconnected	v3	22.1 ms	22.8 ms	25.2 ms	20.8 ms	27
algo: hopcroft-tarjan biconnected	gpu	798 µs	819 µs	1.15 ms	775 µs	834
algo: hierholzer (directed)	v3	4.57 ms	4.82 ms	7.67 ms	4.32 ms	138
algo: hierholzer (directed)	gpu	822 µs	844 µs	1.15 ms	794 µs	807
algo: betweennessCentrality (unweighted)	v3	255 ms	257 ms	266 ms	250 ms	12
algo: betweennessCentrality (unweighted)	gpu	22.4 ms	22.8 ms	24.9 ms	21.8 ms	27
algo: degreeCentralityNormalized	v3	3.79 ms	4.07 ms	7.55 ms	3.66 ms	168
algo: degreeCentralityNormalized	gpu	94.8 µs	98.2 µs	154 µs	91.6 µs	7090
algo: closenessCentrality (one root)	v3	14.7 ms	15.4 ms	18.1 ms	13.9 ms	42
algo: closenessCentrality (one root)	gpu	192 µs	197 µs	282 µs	185 µs	3479
algo: degreeCentrality (one root)	v3	8.15 µs	8.00 µs	8.48 µs	7.28 µs	18
algo: degreeCentrality (one root)	gpu	242 ns	246 ns	301 ns	236 ns	691
algo: pageRank (20 iters)	v3	2.73 ms	2.92 ms	6.31 ms	2.64 ms	229
algo: pageRank (20 iters)	gpu	1.45 ms	1.66 ms	3.82 ms	1.42 ms	418
algo: closenessCentrality (weighted, one root)	v3	14.7 ms	15.7 ms	18.7 ms	14.1 ms	41
algo: closenessCentrality (weighted, one root)	gpu	249 µs	257 µs	414 µs	243 µs	2715
algo: floydWarshall	v3	248 ms	274 ms	388 ms	243 ms	12
algo: floydWarshall	gpu	139 ms	146 ms	142 ms	137 ms	12
algo: closenessCentralityNormalized	v3	274 ms	274 ms	278 ms	270 ms	12
algo: closenessCentralityNormalized	gpu	164 ms	165 ms	168 ms	162 ms	12
algo: markovClustering	v3	4.12 s	4.37 s	4.64 s	4.10 s	12
algo: markovClustering	gpu	3.29 s	3.46 s	3.80 s	3.27 s	12
algo: hierarchicalClustering (threshold)	v3	34.8 ms	34.7 ms	35.7 ms	33.9 ms	16
algo: hierarchicalClustering (threshold)	gpu	15.0 ms	15.3 ms	16.8 ms	14.6 ms	42
algo: kMeans (fixed centroids)	v3	8.97 ms	9.04 ms	9.90 ms	8.78 ms	74
algo: kMeans (fixed centroids)	gpu	2.33 ms	2.38 ms	3.51 ms	2.26 ms	287
algo: kMedoids	v3	363 ms	347 ms	429 ms	226 ms	12
algo: kMedoids	gpu	42.1 ms	44.6 ms	59.4 ms	28.9 ms	12
algo: fuzzyCMeans	v3	79.8 ms	81.3 ms	85.2 ms	78.0 ms	12
algo: fuzzyCMeans	gpu	9.90 ms	9.80 ms	10.9 ms	9.05 ms	67
algo: affinityPropagation	v3	215 ms	214 ms	223 ms	201 ms	12
algo: affinityPropagation	gpu	121 ms	133 ms	177 ms	111 ms	12
algo: kargerStein	v3	316 ms	321 ms	329 ms	313 ms	12
algo: kargerStein	gpu	308 ms	310 ms	317 ms	302 ms	12
algo: betweennessCentrality (weighted)	v3	708 ms	710 ms	721 ms	703 ms	12
algo: betweennessCentrality (weighted)	gpu	80.0 ms	82.8 ms	93.9 ms	76.4 ms	12

mappers N=2,000 · 8.2 s

mapper: bulk write, color — cpu eval

2.68 ms

mapper: bulk write, color — gpu path

177 µs

mapper: bulk write, color + size — cpu eval

3.13 ms

mapper: bulk write, color + size — gpu color / cpu size

2.61 ms

mapper: bulk write, mapped labels (round-5 path) — label refresh

4.07 ms

100 µs1 ms10 ms

All stats (table view)

group	bench	p50	avg	p99	min	samples
mapper: bulk write, color	cpu eval	2.68 ms	2.73 ms	4.12 ms	2.55 ms	250
mapper: bulk write, color	gpu path	177 µs	187 µs	330 µs	172 µs	3755
mapper: bulk write, color + size	cpu eval	3.13 ms	3.19 ms	3.75 ms	2.97 ms	215
mapper: bulk write, color + size	gpu color / cpu size	2.61 ms	2.68 ms	3.52 ms	2.51 ms	258
mapper: bulk write, mapped labels (round-5 path)	label refresh	4.07 ms	4.41 ms	7.29 ms	3.47 ms	154

layouts N=2,000 · 30.9 sv4 (gpu) v3

91×

19×

57×

12×

26×

35×

11×

119 ms

layout: force seed — spectral vs scatter (20 iterations) — spectral (landmark MDS, the default)

112 ms

58.0 ms

371 µs

31.6 µs

526 ns

layout: contract fixed cost (empty impl vs bulk placement) — wrapper + bulk write

39.7 µs

9.5×

100 ns1 µs10 µs100 µs1 ms10 ms100 ms1 s

All stats (table view)

group	bench	p50	avg	p99	min	samples
layout: grid	v3	11.9 ms	12.7 ms	18.2 ms	10.7 ms	50
layout: grid	gpu	130 µs	145 µs	774 µs	82.4 µs	4570
layout: preset	v3	5.59 ms	5.78 ms	7.69 ms	5.40 ms	111
layout: preset	gpu	297 µs	327 µs	908 µs	288 µs	2116
layout: preset (fn form)	v3	5.64 ms	6.02 ms	9.31 ms	5.29 ms	112
layout: preset (fn form)	gpu	98.3 µs	109 µs	375 µs	95.8 µs	6327
layout: circle	v3	9.43 ms	9.86 ms	15.7 ms	8.99 ms	66
layout: circle	gpu	778 µs	877 µs	2.98 ms	749 µs	766
layout: concentric	v3	32.1 ms	32.4 ms	36.0 ms	28.3 ms	17
layout: concentric	gpu	1.23 ms	1.40 ms	4.83 ms	1.16 ms	484
layout: breadthfirst	v3	398 ms	409 ms	441 ms	390 ms	12
layout: breadthfirst	gpu	11.4 ms	11.8 ms	16.1 ms	10.5 ms	51
layout: random	v3	5.53 ms	5.77 ms	9.36 ms	5.33 ms	116
layout: random	gpu	518 µs	571 µs	1.46 ms	494 µs	1208
layout: force (CPU executor, 20 iterations)	gpu	119 ms	131 ms	143 ms	107 ms	12
layout: force seed — spectral vs scatter (20 iterations)	spectral (landmark MDS, the default)	112 ms	111 ms	119 ms	103 ms	12
layout: force seed — spectral vs scatter (20 iterations)	scatter	58.0 ms	60.5 ms	63.4 ms	57.1 ms	12
layout: plumbing (layoutPositions vs setPositions)	finisher	371 µs	415 µs	870 µs	357 µs	1660
layout: plumbing (layoutPositions vs setPositions)	bulk	31.6 µs	44.2 µs	257 µs	29.5 µs	14494
layout: contract fixed cost (empty impl vs bulk placement)	wrapper only	526 ns	984 ns	3.79 µs	514 ns	164
layout: contract fixed cost (empty impl vs bulk placement)	wrapper + bulk write	39.7 µs	41.7 µs	49.6 µs	38.6 µs	12
layout: scope (eles.layout, 10% of nodes)	v3	958 µs	1.07 ms	2.36 ms	928 µs	635
layout: scope (eles.layout, 10% of nodes)	gpu	100 µs	114 µs	275 µs	95.9 µs	6058

style N=2,000 · 27.6 sv4 (gpu) v3

style: sheet swap (compile + applyAll)

1.20×

style: compile vs compile + apply — compile only (in batch)

27.3 µs

style: compile vs compile + apply — compile + apply

10.6 ms

style: first apply on add (256-node band)

1.57×

style: applyAll, flat vs compound (100 parents — the 14.6 partition) — flat

10.5 ms

style: applyAll, flat vs compound (100 parents — the 14.6 partition) — compound

11.1 ms

style: applyAll — constant vs state-case (the 57.1d partition) — constant sheet

10.4 ms

style: applyAll — constant vs state-case (the 57.1d partition) — state-case sheet

10.3 ms

style: applyAll — state-case vs data-case (what the partition avoids) — state-case sheet

10.5 ms

style: applyAll — state-case vs data-case (what the partition avoids) — data-case sheet

14.0 ms

style: select+unselect 256-band — constant vs state-case sheet — constant sheet (skips restyle)

6.76 µs

style: select+unselect 256-band — constant vs state-case sheet — paint-only state sheet (round-61 diff path)

34.7 µs

style: select+unselect 256-band — constant vs state-case sheet — state-case sheet (all channels; full write)

611 µs

style: read one prop (color)

1.90×

style: read one prop (width)

2.1×

style: numericStyle

2.8×

style: renderedStyle (one prop)

14×

style: whole-object style()

6.2×

10 ns100 ns1 µs10 µs100 µs1 ms10 ms100 ms

All stats (table view)

group	bench	p50	avg	p99	min	samples
style: sheet swap (compile + applyAll)	v3	13.1 ms	13.4 ms	17.0 ms	12.1 ms	46
style: sheet swap (compile + applyAll)	gpu	10.8 ms	10.9 ms	11.7 ms	10.6 ms	57
style: compile vs compile + apply	compile only (in batch)	27.3 µs	34.8 µs	69.9 µs	23.5 µs	19044
style: compile vs compile + apply	compile + apply	10.6 ms	10.7 ms	13.1 ms	10.3 ms	62
style: first apply on add (256-node band)	v3	1.31 ms	1.56 ms	7.13 ms	1.11 ms	440
style: first apply on add (256-node band)	gpu	833 µs	921 µs	2.68 ms	805 µs	736
style: applyAll, flat vs compound (100 parents — the 14.6 partition)	flat	10.5 ms	10.6 ms	12.7 ms	10.2 ms	62
style: applyAll, flat vs compound (100 parents — the 14.6 partition)	compound	11.1 ms	11.5 ms	15.0 ms	10.8 ms	57
style: applyAll — constant vs state-case (the 57.1d partition)	constant sheet	10.4 ms	10.4 ms	11.0 ms	10.2 ms	60
style: applyAll — constant vs state-case (the 57.1d partition)	state-case sheet	10.3 ms	10.4 ms	10.7 ms	10.2 ms	62
style: applyAll — state-case vs data-case (what the partition avoids)	state-case sheet	10.5 ms	10.7 ms	12.4 ms	10.2 ms	62
style: applyAll — state-case vs data-case (what the partition avoids)	data-case sheet	14.0 ms	14.1 ms	15.4 ms	13.6 ms	43
style: select+unselect 256-band — constant vs state-case sheet	constant sheet (skips restyle)	6.76 µs	6.80 µs	6.95 µs	6.68 µs	20
style: select+unselect 256-band — constant vs state-case sheet	paint-only state sheet (round-61 diff path)	34.7 µs	36.1 µs	65.9 µs	33.3 µs	18530
style: select+unselect 256-band — constant vs state-case sheet	state-case sheet (all channels; full write)	611 µs	636 µs	1.04 ms	590 µs	1093
style: read one prop (color)	v3	108 ns	120 ns	598 ns	96.0 ns	1994138
style: read one prop (color)	gpu	56.9 ns	59.7 ns	79.6 ns	54.6 ns	2856
style: read one prop (width)	v3	111 ns	115 ns	442 ns	96.0 ns	2005941
style: read one prop (width)	gpu	53.0 ns	56.5 ns	114 ns	50.7 ns	3005
style: numericStyle	v3	138 ns	127 ns	348 ns	85.0 ns	1842844
style: numericStyle	gpu	49.8 ns	52.4 ns	116 ns	48.4 ns	3265
style: renderedStyle (one prop)	v3	913 ns	971 ns	1.97 µs	880 ns	543032
style: renderedStyle (one prop)	gpu	63.4 ns	66.8 ns	211 ns	61.5 ns	2567
style: whole-object style()	v3	87.0 µs	92.6 µs	177 µs	84.6 µs	7522
style: whole-object style()	gpu	14.0 µs	15.6 µs	49.6 µs	13.5 µs	43314

load N=2,000 · 19.5 sv4 (gpu) v3

load: init from the definition form (construct + dispose)

5.9×

load: v4 ingest forms (construct + dispose) — definition form

21.7 ms

load: v4 ingest forms (construct + dispose) — columnar form

12.6 ms

load: v4 ingest forms (construct + dispose) — wire buffer

12.1 ms

load: v4 ingest forms (construct + dispose) — def clone alone (control)

308 µs

load: conversion — toColumnarElements

709 µs

load: conversion — serializeElements (from columnar)

206 µs

load: conversion — serializeElements (from defs)

891 µs

load: conversion — deserializeElements

4.55 µs

load: export from a live graph — cy.serialize() (wire)

493 µs

load: export from a live graph — cy.json() (v4)

730 µs

load: cy.json()

2.7×

load: add a 256-node band (add + remove)

1.36×

load: add a 256-node band, v4 forms — definition form

2.63 ms

load: add a 256-node band, v4 forms — columnar form

2.24 ms

load: add a 256-node band, v4 forms — wire buffer

2.20 ms

1 µs10 µs100 µs1 ms10 ms100 ms1 s

All stats (table view)

group	bench	p50	avg	p99	min	samples
load: init from the definition form (construct + dispose)	v3	134 ms	144 ms	172 ms	126 ms	12
load: init from the definition form (construct + dispose)	gpu	22.5 ms	23.5 ms	28.8 ms	20.1 ms	25
load: v4 ingest forms (construct + dispose)	definition form	21.7 ms	22.0 ms	26.0 ms	18.9 ms	27
load: v4 ingest forms (construct + dispose)	columnar form	12.6 ms	13.4 ms	23.1 ms	12.0 ms	47
load: v4 ingest forms (construct + dispose)	wire buffer	12.1 ms	13.2 ms	25.0 ms	11.8 ms	47
load: v4 ingest forms (construct + dispose)	def clone alone (control)	308 µs	632 µs	11.1 ms	218 µs	1040
load: conversion	toColumnarElements	709 µs	774 µs	1.46 ms	685 µs	888
load: conversion	serializeElements (from columnar)	206 µs	376 µs	1.05 ms	173 µs	1836
load: conversion	serializeElements (from defs)	891 µs	947 µs	2.02 ms	866 µs	734
load: conversion	deserializeElements	4.55 µs	4.30 µs	5.76 µs	3.29 µs	30
load: export from a live graph	cy.serialize() (wire)	493 µs	565 µs	1.92 ms	485 µs	1217
load: export from a live graph	cy.json() (v4)	730 µs	1.07 ms	6.75 ms	696 µs	618
load: cy.json()	v3	2.00 ms	2.32 ms	5.14 ms	1.95 ms	293
load: cy.json()	gpu	746 µs	827 µs	2.78 ms	727 µs	842
load: add a 256-node band (add + remove)	v3	3.65 ms	4.26 ms	10.7 ms	3.17 ms	154
load: add a 256-node band (add + remove)	gpu	2.68 ms	2.96 ms	6.58 ms	2.07 ms	227
load: add a 256-node band, v4 forms	definition form	2.63 ms	2.66 ms	4.13 ms	2.06 ms	256
load: add a 256-node band, v4 forms	columnar form	2.24 ms	2.31 ms	3.21 ms	1.83 ms	292
load: add a 256-node band, v4 forms	wire buffer	2.20 ms	2.27 ms	3.35 ms	1.83 ms	302

spatial N=2,000 · 37.5 sv4 (gpu) v3

pick: hit (the walk; one inside-test) — zoom 1

20.6 µs

pick: hit (the walk; one inside-test) — zoom 2

21.0 µs

pick: miss walk — exact vs 24 px touch halo (57.9) — exact

46.2 µs

pick: miss walk — exact vs 24 px touch halo (57.9) — padPx 24

45.6 µs

pick: shape inside-test, 2000 coincident misses — ellipse

87.2 µs

pick: shape inside-test, 2000 coincident misses — star

235 µs

pick: shape inside-test, 2000 coincident misses — round-hexagon

793 µs

pick: shape inside-test, 2000 coincident misses — barrel

681 µs

pick: shape inside-test, 2000 coincident misses — cut-rectangle

99.8 µs

pick: shape inside-test, 2000 coincident misses — polygon

214 µs

pick: round-hexagon inside-test by zoom (not affine-invariant) — zoom 1

802 µs

pick: round-hexagon inside-test by zoom (not affine-invariant) — zoom 2

799 µs

pick: background miss (full highWater walk) — gpu

45.1 µs

pick: label box in the scan (text-events) — text-events: no

45.3 µs

pick: label box in the scan (text-events) — text-events: yes

127 µs

box: elementsInBox by covered fraction — 10%

149 µs

box: elementsInBox by covered fraction — 50%

192 µs

box: elementsInBox by covered fraction — 100%

324 µs

box: label containment (boxSelectionIncludesLabels) — off

230 µs

box: label containment (boxSelectionIncludesLabels) — on

264 µs

box: overlap vs contain (straight edges) — contain

235 µs

box: overlap vs contain (straight edges) — overlap

475 µs

box: overlap vs contain (curved edges — the segment walk) — contain

1.26 ms

box: overlap vs contain (curved edges — the segment walk) — overlap

2.14 ms

bounds: whole-graph boundingBox()

6.1×

bounds: boundingBox, labels in vs out (v4) — includeLabels: true (default)

1.25 ms

bounds: boundingBox, labels in vs out (v4) — includeLabels: false

715 µs

bounds: cy.fit() (whole-graph scan)

33×

bounds: getFitViewport() (compute without committing)

33×

bounds: boundingBoxAt (v4 only) — gpu

2.53 ms

bounds: one node boundingBox()

1.75×

bounds: labelBoundingBox (v4 only) — gpu

409 ns

bounds: compound + taxi conservative scan (round 54, v4 only) — warm (curve bbs memoized)

75.2 µs

bounds: compound + taxi conservative scan (round 54, v4 only) — cold (geometry write per call)

152 µs

100 ns1 µs10 µs100 µs1 ms10 ms

All stats (table view)

group	bench	p50	avg	p99	min	samples
pick: hit (the walk; one inside-test)	zoom 1	20.6 µs	21.5 µs	41.6 µs	20.5 µs	31955
pick: hit (the walk; one inside-test)	zoom 2	21.0 µs	21.4 µs	21.8 µs	20.8 µs	12
pick: miss walk — exact vs 24 px touch halo (57.9)	exact	46.2 µs	46.7 µs	48.5 µs	45.8 µs	12
pick: miss walk — exact vs 24 px touch halo (57.9)	padPx 24	45.6 µs	47.8 µs	95.1 µs	43.1 µs	14606
pick: shape inside-test, 2000 coincident misses	ellipse	87.2 µs	90.7 µs	160 µs	85.3 µs	7696
pick: shape inside-test, 2000 coincident misses	star	235 µs	246 µs	340 µs	225 µs	2835
pick: shape inside-test, 2000 coincident misses	round-hexagon	793 µs	809 µs	1.05 ms	784 µs	865
pick: shape inside-test, 2000 coincident misses	barrel	681 µs	712 µs	1.30 ms	658 µs	978
pick: shape inside-test, 2000 coincident misses	cut-rectangle	99.8 µs	103 µs	176 µs	95.4 µs	6766
pick: shape inside-test, 2000 coincident misses	polygon	214 µs	217 µs	287 µs	207 µs	3232
pick: round-hexagon inside-test by zoom (not affine-invariant)	zoom 1	802 µs	829 µs	1.37 ms	788 µs	844
pick: round-hexagon inside-test by zoom (not affine-invariant)	zoom 2	799 µs	809 µs	1.01 ms	780 µs	862
pick: background miss (full highWater walk)	gpu	45.1 µs	45.6 µs	46.1 µs	44.9 µs	12
pick: label box in the scan (text-events)	text-events: no	45.3 µs	45.7 µs	46.3 µs	44.8 µs	12
pick: label box in the scan (text-events)	text-events: yes	127 µs	131 µs	262 µs	123 µs	5281
box: elementsInBox by covered fraction	10%	149 µs	156 µs	308 µs	147 µs	4495
box: elementsInBox by covered fraction	50%	192 µs	199 µs	293 µs	191 µs	3492
box: elementsInBox by covered fraction	100%	324 µs	337 µs	785 µs	316 µs	2068
box: label containment (boxSelectionIncludesLabels)	off	230 µs	238 µs	455 µs	224 µs	2931
box: label containment (boxSelectionIncludesLabels)	on	264 µs	279 µs	621 µs	259 µs	2504
box: overlap vs contain (straight edges)	contain	235 µs	244 µs	406 µs	230 µs	2863
box: overlap vs contain (straight edges)	overlap	475 µs	486 µs	901 µs	418 µs	1433
box: overlap vs contain (curved edges — the segment walk)	contain	1.26 ms	1.28 ms	1.95 ms	1.22 ms	541
box: overlap vs contain (curved edges — the segment walk)	overlap	2.14 ms	2.20 ms	2.96 ms	2.11 ms	315
bounds: whole-graph boundingBox()	v3	7.64 ms	7.77 ms	8.93 ms	7.34 ms	85
bounds: whole-graph boundingBox()	gpu	1.25 ms	1.29 ms	2.22 ms	1.22 ms	539
bounds: boundingBox, labels in vs out (v4)	includeLabels: true (default)	1.25 ms	1.28 ms	1.85 ms	1.22 ms	543
bounds: boundingBox, labels in vs out (v4)	includeLabels: false	715 µs	739 µs	1.43 ms	683 µs	943
bounds: cy.fit() (whole-graph scan)	v3	7.75 ms	7.87 ms	9.00 ms	7.52 ms	85
bounds: cy.fit() (whole-graph scan)	gpu	238 µs	247 µs	442 µs	232 µs	2828
bounds: getFitViewport() (compute without committing)	v3	7.85 ms	8.08 ms	10.8 ms	7.58 ms	82
bounds: getFitViewport() (compute without committing)	gpu	235 µs	245 µs	444 µs	228 µs	2850
bounds: boundingBoxAt (v4 only)	gpu	2.53 ms	2.64 ms	3.64 ms	2.52 ms	257
bounds: one node boundingBox()	v3	1.74 µs	1.75 µs	1.89 µs	1.68 µs	92
bounds: one node boundingBox()	gpu	995 ns	1.03 µs	1.46 µs	961 ns	160
bounds: labelBoundingBox (v4 only)	gpu	409 ns	439 ns	1.07 µs	405 ns	382
bounds: compound + taxi conservative scan (round 54, v4 only)	warm (curve bbs memoized)	75.2 µs	83.2 µs	186 µs	74.3 µs	8336
bounds: compound + taxi conservative scan (round 54, v4 only)	cold (geometry write per call)	152 µs	160 µs	232 µs	149 µs	4349

data N=2,000 · 28.3 sv4 (gpu) v3

data: bulk write, numeric column

21×

data: bulk write, dictionary string column (4 values)

17×

data: bulk write, one new dictionary entry per pass

17×

data: bulk write, object column (plain-array fallback)

20×

data: removeData (whole collection)

7.7×

data: read one numeric key

1.36×

data: read one string key (dictionary decode)

1.33×

data: read the whole object

2.3×

query: data equality ([bucket = 3] vs { data: { bucket: 3 } })

13×

query: data comparison ([weight > 3] vs { gt })

12×

query: two keys AND-ed

12×

query: membership ([bucket in ...] vs { in })

15×

query: predicate function (both sides)

6.7×

query: structural, parents (:parent vs { parent: true })

33×

query: structural, childless (:childless vs { parent: false })

13×

query: structural, orphans (:orphan vs { child: false })

33×

10 ns100 ns1 µs10 µs100 µs1 ms10 ms

All stats (table view)

group	bench	p50	avg	p99	min	samples
data: bulk write, numeric column	v3	1.80 ms	1.85 ms	2.95 ms	1.73 ms	362
data: bulk write, numeric column	gpu	84.3 µs	87.1 µs	128 µs	81.3 µs	8027
data: bulk write, dictionary string column (4 values)	v3	1.68 ms	1.78 ms	3.69 ms	1.64 ms	375
data: bulk write, dictionary string column (4 values)	gpu	101 µs	106 µs	316 µs	94.8 µs	6592
data: bulk write, one new dictionary entry per pass	v3	1.66 ms	1.75 ms	2.88 ms	1.65 ms	390
data: bulk write, one new dictionary entry per pass	gpu	99.8 µs	105 µs	189 µs	96.8 µs	6698
data: bulk write, object column (plain-array fallback)	v3	1.68 ms	1.74 ms	2.90 ms	1.64 ms	398
data: bulk write, object column (plain-array fallback)	gpu	85.5 µs	91.9 µs	130 µs	82.9 µs	7600
data: removeData (whole collection)	v3	2.52 ms	2.66 ms	4.47 ms	2.38 ms	257
data: removeData (whole collection)	gpu	328 µs	348 µs	906 µs	322 µs	1984
data: read one numeric key	v3	38.5 ns	40.4 ns	119 ns	37.1 ns	4250
data: read one numeric key	gpu	28.3 ns	29.6 ns	76.2 ns	27.8 ns	5596
data: read one string key (dictionary decode)	v3	41.8 ns	43.5 ns	88.1 ns	40.5 ns	3943
data: read one string key (dictionary decode)	gpu	31.4 ns	32.8 ns	68.5 ns	27.4 ns	5229
data: read the whole object	v3	48.0 ns	49.8 ns	80.8 ns	45.9 ns	3449
data: read the whole object	gpu	20.9 ns	22.2 ns	43.3 ns	20.3 ns	7680
query: data equality ([bucket = 3] vs { data: { bucket: 3 } })	v3	661 µs	688 µs	1.12 ms	641 µs	1014
query: data equality ([bucket = 3] vs { data: { bucket: 3 } })	gpu	52.3 µs	56.5 µs	121 µs	42.9 µs	12138
query: data comparison ([weight > 3] vs { gt })	v3	738 µs	757 µs	1.08 ms	730 µs	920
query: data comparison ([weight > 3] vs { gt })	gpu	59.6 µs	67.7 µs	180 µs	57.1 µs	9942
query: two keys AND-ed	v3	777 µs	807 µs	1.25 ms	768 µs	865
query: two keys AND-ed	gpu	67.0 µs	69.9 µs	121 µs	64.0 µs	9687
query: membership ([bucket in ...] vs { in })	v3	1.01 ms	1.04 ms	1.41 ms	976 µs	660
query: membership ([bucket in ...] vs { in })	gpu	65.3 µs	71.8 µs	163 µs	63.6 µs	9385
query: predicate function (both sides)	v3	657 µs	672 µs	911 µs	638 µs	1040
query: predicate function (both sides)	gpu	98.5 µs	103 µs	179 µs	95.7 µs	6754
query: structural, parents (:parent vs { parent: true })	v3	516 µs	525 µs	657 µs	499 µs	1332
query: structural, parents (:parent vs { parent: true })	gpu	15.6 µs	17.9 µs	23.1 µs	15.1 µs	12
query: structural, childless (:childless vs { parent: false })	v3	798 µs	818 µs	1.30 ms	765 µs	852
query: structural, childless (:childless vs { parent: false })	gpu	60.5 µs	68.6 µs	188 µs	56.3 µs	9772
query: structural, orphans (:orphan vs { child: false })	v3	503 µs	511 µs	604 µs	491 µs	1365
query: structural, orphans (:orphan vs { child: false })	gpu	15.5 µs	16.3 µs	17.9 µs	15.1 µs	12

events N=2,000 · 27.0 sv4 (gpu) v3

emit: position write, no listeners

29×

emit: 1 core listener

8.5×

emit: 1 element (ref-qualified) listener

8.3×

emit: delegated (v3 selector vs v4 predicate)

8.9×

emit: 10 core listeners

6.9×

emit: on() + off() round-trip

2.2×

emit: flat vs phased (compound bubbling, 14.5) — orphan (flat emit)

307 ns

emit: flat vs phased (compound bubbling, 14.5) — child of 2 ancestors (phased)

915 ns

emit: flat vs phased, no listeners — orphan (flat)

76.2 ns

emit: flat vs phased, no listeners — child (phased)

468 ns

anim: start + stop, one element

1.37×

anim: start against a running animation (round-21 eviction) — disjoint channels (composes)

1.28 µs

anim: start against a running animation (round-21 eviction) — overlapping channels (evicts)

810 ns

anim: start + stop, 512-node collection

3.0×

anim: delay() (no channels touched)

1.58×

anim: controls (v4 only) — pause + resume

452 ns

anim: controls (v4 only) — reverse

903 ns

anim: viewport animate + stop

2.8×

10 ns100 ns1 µs10 µs100 µs1 ms

All stats (table view)

group	bench	p50	avg	p99	min	samples
emit: position write, no listeners	v3	1.31 µs	1.39 µs	1.69 µs	1.27 µs	114
emit: position write, no listeners	gpu	45.7 ns	48.4 ns	86.8 ns	44.6 ns	3250
emit: 1 core listener	v3	1.35 µs	1.39 µs	2.05 µs	1.31 µs	119
emit: 1 core listener	gpu	158 ns	163 ns	284 ns	154 ns	962
emit: 1 element (ref-qualified) listener	v3	1.38 µs	1.46 µs	2.26 µs	1.34 µs	106
emit: 1 element (ref-qualified) listener	gpu	167 ns	172 ns	257 ns	165 ns	923
emit: delegated (v3 selector vs v4 predicate)	v3	1.52 µs	1.55 µs	2.16 µs	1.47 µs	101
emit: delegated (v3 selector vs v4 predicate)	gpu	169 ns	179 ns	336 ns	164 ns	898
emit: 10 core listeners	v3	1.92 µs	1.98 µs	2.30 µs	1.87 µs	79
emit: 10 core listeners	gpu	279 ns	289 ns	432 ns	277 ns	559
emit: on() + off() round-trip	v3	477 ns	587 ns	1.48 µs	436 ns	278
emit: on() + off() round-trip	gpu	216 ns	262 ns	932 ns	184 ns	628
emit: flat vs phased (compound bubbling, 14.5)	orphan (flat emit)	307 ns	357 ns	2.12 µs	280 ns	955537
emit: flat vs phased (compound bubbling, 14.5)	child of 2 ancestors (phased)	915 ns	968 ns	1.31 µs	834 ns	163
emit: flat vs phased, no listeners	orphan (flat)	76.2 ns	84.2 ns	169 ns	72.5 ns	1970
emit: flat vs phased, no listeners	child (phased)	468 ns	515 ns	1.59 µs	456 ns	324
anim: start + stop, one element	v3	892 ns	1.33 µs	3.58 µs	814 ns	389399
anim: start + stop, one element	gpu	650 ns	1.10 µs	4.67 µs	604 ns	453288
anim: start against a running animation (round-21 eviction)	disjoint channels (composes)	1.28 µs	1.92 µs	3.88 µs	1.08 µs	81
anim: start against a running animation (round-21 eviction)	overlapping channels (evicts)	810 ns	1.39 µs	3.58 µs	795 ns	110
anim: start + stop, 512-node collection	v3	299 µs	432 µs	1.25 ms	288 µs	1443
anim: start + stop, 512-node collection	gpu	99.2 µs	111 µs	168 µs	95.2 µs	5762
anim: delay() (no channels touched)	v3	656 ns	957 ns	3.33 µs	641 ns	167
anim: delay() (no channels touched)	gpu	414 ns	729 ns	2.73 µs	410 ns	225
anim: controls (v4 only)	pause + resume	452 ns	729 ns	2.64 µs	447 ns	210
anim: controls (v4 only)	reverse	903 ns	1.53 µs	3.53 µs	879 ns	100
anim: viewport animate + stop	v3	558 ns	855 ns	2.91 µs	551 ns	192
anim: viewport animate + stop	gpu	201 ns	238 ns	1.41 µs	195 ns	700

store N=2,000 · 20.9 s

id index: set x 2000 (build) — gpu

283 µs

id index: single-key operations — get (probe)

57.5 ns

id index: single-key operations — has

51.3 ns

id index: single-key operations — idAt (memoized hit)

9.02 ns

id index: single-key operations — hashAt

7.01 ns

id index: remove + re-set (churn, drives the round-11 reclaim) — gpu

232 ns

adjacency: rebuild (4000 edges, two counting passes) — gpu

56.2 µs

adjacency: reads — outDegree (O(1))

6.19 ns

adjacency: reads — outEdges

45.0 ns

adjacency: reads — connectedEdges

124 ns

adjacency: addEdge + removeEdge (the overlay path) — gpu

137 ns

blob: write x 2000 (build) — gpu

103 µs

blob: single-record operations — rewrite in place (same length)

29.9 ns

blob: single-record operations — offsetOf

4.25 ns

blob: single-record operations — free + rewrite

149 ns

dirty: mark (per-write span coalescing) — contiguous marks

12.8 ns

dirty: mark (per-write span coalescing) — scattered marks

17.8 ns

dirty: take (a frame's drain) — gpu

598 ns

images: registry bookkeeping — acquire (existing url)

201 ns

images: registry bookkeeping — acquire (new url) + release

534 ns

charts: sheet apply, chart vs no chart — no chart

10.4 ms

charts: sheet apply, chart vs no chart — pie from data (blob record per node)

10.9 ms

charts: data write refreshing chart-values — gpu

2.43 ms

1 ns10 ns100 ns1 µs10 µs100 µs1 ms10 ms100 ms

All stats (table view)

group	bench	p50	avg	p99	min	samples
id index: set x 2000 (build)	gpu	283 µs	296 µs	467 µs	272 µs	2368
id index: single-key operations	get (probe)	57.5 ns	60.0 ns	93.6 ns	56.2 ns	2858
id index: single-key operations	has	51.3 ns	52.5 ns	74.0 ns	49.5 ns	3257
id index: single-key operations	idAt (memoized hit)	9.02 ns	9.33 ns	17.6 ns	8.63 ns	18230
id index: single-key operations	hashAt	7.01 ns	7.22 ns	13.6 ns	6.89 ns	23453
id index: remove + re-set (churn, drives the round-11 reclaim)	gpu	232 ns	241 ns	347 ns	226 ns	695
adjacency: rebuild (4000 edges, two counting passes)	gpu	56.2 µs	64.4 µs	252 µs	54.8 µs	10786
adjacency: reads	outDegree (O(1))	6.19 ns	6.28 ns	8.68 ns	5.90 ns	27059
adjacency: reads	outEdges	45.0 ns	46.7 ns	63.4 ns	43.3 ns	3647
adjacency: reads	connectedEdges	124 ns	188 ns	2.20 µs	117 ns	890
adjacency: addEdge + removeEdge (the overlay path)	gpu	137 ns	141 ns	248 ns	135 ns	1194
blob: write x 2000 (build)	gpu	103 µs	127 µs	790 µs	99.8 µs	5469
blob: single-record operations	rewrite in place (same length)	29.9 ns	30.7 ns	45.4 ns	29.0 ns	5565
blob: single-record operations	offsetOf	4.25 ns	4.38 ns	7.99 ns	4.09 ns	38412
blob: single-record operations	free + rewrite	149 ns	178 ns	629 ns	146 ns	950
dirty: mark (per-write span coalescing)	contiguous marks	12.8 ns	13.2 ns	23.0 ns	12.1 ns	12919
dirty: mark (per-write span coalescing)	scattered marks	17.8 ns	18.3 ns	28.2 ns	17.1 ns	9322
dirty: take (a frame's drain)	gpu	598 ns	703 ns	1.32 µs	583 ns	236
images: registry bookkeeping	acquire (existing url)	201 ns	208 ns	277 ns	199 ns	820
images: registry bookkeeping	acquire (new url) + release	534 ns	616 ns	2.79 µs	521 ns	267
charts: sheet apply, chart vs no chart	no chart	10.4 ms	10.6 ms	12.3 ms	10.1 ms	61
charts: sheet apply, chart vs no chart	pie from data (blob record per node)	10.9 ms	11.0 ms	12.0 ms	10.6 ms	59
charts: data write refreshing chart-values	gpu	2.43 ms	2.49 ms	3.54 ms	2.34 ms	277

surface N=2,000 · 2.9 minv4 (gpu) v3

viewport: zoom() get

4.7×

viewport: zoom( n ) set

1.80×

viewport: pan() get

1.08×

viewport: pan( p ) set

1.95×

viewport: panBy()

1.73×

viewport: extent()

1.58×

viewport: renderedExtent()

1.65×

viewport: width() + height() + size() (3 calls)

2.9×

viewport: minZoom/maxZoom get (2 calls)

15×

viewport: center( eles )

19×

viewport: reset()

121×

viewport: viewport({ zoom, pan })

64×

viewport: getCenterPan( eles ) — gpu

6.26 µs

core: instanceString + headless + styleEnabled (3 calls)

1.25×

core: hasElementWithId + hasCompoundNodes (2 calls)

1.27×

core: options()

86×

core: mutableElements() (v4 materializes; v3 is O(1))

1.36×

core: gating getters (5 calls)

1.16×

core: gating setters round-trip (autolock/autoungrabify/autounselectify)

1.85×

core: interaction tuning getters (5 calls) — gpu

7.15 ns

core: graph data() set + get (v3 restyles the graph)

8610×

core: graph scratch() set + get (v3 restyles the graph)

16571×

core: batch( fn ) (empty)

6.8×

core: startBatch + endBatch + batching

5.8×

core: collection()

7.2×

core: compact() (no dead slots — the early-out) — gpu

16.9 µs

collection: element accessors (id/group/isNode/isEdge, 4 calls)

1.79×

collection: length + empty + nonempty (3 calls)

4.5×

collection: eq/first/last (3 calls)

1.83×

collection: slice()

2.1×

collection: toArray()

12×

collection: sort( fn )

1.57×

collection: reduce( fn )

1.38×

collection: min + max ( fn ) (2 calls)

2.0×

collection: some + every (2 calls)

1.57×

collection: same + anySame + contains (3 calls)

72×

collection: allAre (all match: the full walk)

1.52×

collection: is (none match: the full walk)

2.2×

collection: union

6.3×

collection: intersection

27×

collection: difference

28×

collection: symmetricDifference

3.7×

collection: absoluteComplement

13×

collection: byGroup

2.6×

collection: diff

2.9×

collection: indexOf

1.41×

traversal: openNeighborhood + closedNeighborhood (2 calls)

2.9×

traversal: connectedNodes

3.3×

traversal: sources + targets (2 calls)

3.4×

traversal: source() + target() (2 calls)

1.27×

traversal: edgesWith

1.21×

traversal: edgesTo

1.17×

traversal: parallelEdges + codirectedEdges (2 calls)

3.3×

traversal: roots + leaves (2 calls)

15×

traversal: allAreNeighbors

5.1×

traversal: isLoop + isSimple (2 calls)

2.3×

degree: indegree + outdegree (2 calls)

72×

degree: minDegree + maxDegree (2 calls)

35×

degree: minIndegree + maxOutdegree (2 calls)

21×

dimensions: renderedPosition get

2.5×

dimensions: relativePosition get

12×

dimensions: width + height + outerWidth + outerHeight (4 calls)

2.5×

dimensions: rendered dims (4 calls)

2.9×

dimensions: renderedBoundingBox

1.91×

dimensions: midpoint + renderedMidpoint (2 calls) — gpu

1.10 µs

dimensions: source/targetEndpoint (2 calls) — gpu

381 ns

dimensions: rendered source/targetEndpoint (2 calls) — gpu

393 ns

dimensions: shift + silentShift (round-trip)

60×

dimensions: silentPosition

17×

dimensions: positions( fn ) over 100

55×

dimensions: padding + paddedWidth + paddedHeight (3 calls)

15×

curves: isBundledBezier — gpu

20.6 ns

curves: controlPoints + renderedControlPoints (2 calls) — gpu

525 ns

curves: segmentPoints (undefined for bezier) — gpu

33.2 ns

compound: parent + parents + ancestors (3 calls)

3.6×

compound: children + descendants + siblings (3 calls)

4.6×

compound: orphans + nonorphans (2 calls)

6.6×

compound: commonAncestors

131×

compound: isParent/isChild/isChildless/isOrphan (4 calls)

1.57×

flags: select + unselect + selected (band)

15×

flags: selectify + unselectify + selectable

95×

flags: lock + unlock + locked

92×

flags: grabify + ungrabify + grabbable

94×

flags: show + hide + visible

224×

flags: active/activate/unactivate (3 calls)

15×

flags: panify + unpanify + pannable

95×

flags: takesUpSpace + interactive + transparent (3 calls)

2.9×

flags: effectiveOpacity

2.1×

element data: scratch get + set

7.3×

element data: json() one element

1.98×

element data: jsons() over 100

1.87×

element data: label() — gpu

15.4 ns

core: batchData()

8.6×

core: animated()

4.9×

core: zoomRange get — gpu

2.16 ns

core: forceRender() (headless no-op)

16×

core: resize() / invalidateSize()

2.7×

core: onRender + offRender

2.4×

core: removeScratch()

9678×

core: selectionType get

1.99×

core: isReady()

2.2×

core: window()

2.8×

core: container()

2.7×

core: destroyed()

2.2×

core: removeAllListeners()

8.8×

collection: cy()

3.5×

collection: indexOfId()

1.87×

collection: removed() + inside() (2 calls)

1.11×

collection: silentPositions( fn ) over 100

35×

collection: animated()

2.9×

collection: delayAnimation()

2.7×

collection: removeScratch()

9.0×

collection: component()

42×

collection: layoutDimensions()

3.6×

degree: maxIndegree + minOutdegree (2 calls)

22×

flags: grabbed + hidden + inactive (3 calls)

2.0×

curves: renderedSegmentPoints — gpu

36.1 ns

layout contract: edgeSlots + endpoints + degreeOf — gpu

34.7 µs

formats: isColumnarElements + isSerializedElements + isPackedIds — gpu

4.61 ns

0.100 ns1 ns10 ns100 ns1 µs10 µs100 µs1 ms10 ms100 ms

All stats (table view)

group	bench	p50	avg	p99	min	samples
viewport: zoom() get	v3	5.70 ns	6.00 ns	13.8 ns	5.41 ns	28076
viewport: zoom() get	gpu	1.21 ns	1.25 ns	2.20 ns	1.18 ns	126849
viewport: zoom( n ) set	v3	620 ns	628 ns	755 ns	594 ns	264
viewport: zoom( n ) set	gpu	344 ns	360 ns	790 ns	316 ns	463
viewport: pan() get	v3	5.27 ns	5.42 ns	9.26 ns	5.17 ns	31112
viewport: pan() get	gpu	4.86 ns	5.08 ns	11.6 ns	4.85 ns	33363
viewport: pan( p ) set	v3	661 ns	669 ns	745 ns	640 ns	249
viewport: pan( p ) set	gpu	339 ns	343 ns	404 ns	316 ns	484
viewport: panBy()	v3	615 ns	619 ns	703 ns	591 ns	270
viewport: panBy()	gpu	356 ns	366 ns	660 ns	315 ns	458
viewport: extent()	v3	30.6 ns	47.9 ns	197 ns	28.6 ns	3515
viewport: extent()	gpu	19.3 ns	25.9 ns	107 ns	18.1 ns	6493
viewport: renderedExtent()	v3	27.2 ns	28.5 ns	50.6 ns	26.1 ns	5943
viewport: renderedExtent()	gpu	16.5 ns	23.3 ns	99.8 ns	15.2 ns	7233
viewport: width() + height() + size() (3 calls)	v3	20.9 ns	21.9 ns	44.6 ns	20.4 ns	7769
viewport: width() + height() + size() (3 calls)	gpu	7.27 ns	7.72 ns	17.1 ns	6.99 ns	21867
viewport: minZoom/maxZoom get (2 calls)	v3	14.7 ns	15.3 ns	25.2 ns	14.2 ns	11133
viewport: minZoom/maxZoom get (2 calls)	gpu	0.999 ns	1.04 ns	1.46 ns	0.998 ns	151093
viewport: center( eles )	v3	121 µs	126 µs	204 µs	114 µs	5532
viewport: center( eles )	gpu	6.19 µs	6.20 µs	6.29 µs	6.11 µs	23
viewport: reset()	v3	1.07 µs	1.15 µs	1.73 µs	1.00 µs	144
viewport: reset()	gpu	8.83 ns	9.22 ns	16.7 ns	8.57 ns	18301
viewport: viewport({ zoom, pan })	v3	1.01 µs	1.11 µs	1.73 µs	983 ns	148
viewport: viewport({ zoom, pan })	gpu	15.8 ns	16.7 ns	37.9 ns	15.1 ns	10210
viewport: getCenterPan( eles )	gpu	6.26 µs	6.28 µs	6.48 µs	6.12 µs	23
core: instanceString + headless + styleEnabled (3 calls)	v3	32.9 ns	34.4 ns	68.7 ns	32.2 ns	4963
core: instanceString + headless + styleEnabled (3 calls)	gpu	26.3 ns	27.1 ns	37.8 ns	25.2 ns	6291
core: hasElementWithId + hasCompoundNodes (2 calls)	v3	40.5 ns	41.7 ns	52.8 ns	39.8 ns	4093
core: hasElementWithId + hasCompoundNodes (2 calls)	gpu	31.8 ns	33.0 ns	50.9 ns	30.8 ns	5173
core: options()	v3	91.5 ns	94.9 ns	138 ns	90.0 ns	1797
core: options()	gpu	1.07 ns	1.15 ns	2.20 ns	1.05 ns	138738
core: mutableElements() (v4 materializes; v3 is O(1))	v3	3.19 ns	3.34 ns	4.42 ns	3.14 ns	49947
core: mutableElements() (v4 materializes; v3 is O(1))	gpu	2.35 ns	2.43 ns	4.33 ns	2.25 ns	68076
core: gating getters (5 calls)	v3	15.7 ns	16.4 ns	30.2 ns	15.3 ns	10386
core: gating getters (5 calls)	gpu	13.5 ns	14.3 ns	37.5 ns	12.6 ns	11896
core: gating setters round-trip (autolock/autoungrabify/autounselectify)	v3	11.2 ns	11.6 ns	20.2 ns	10.6 ns	14725
core: gating setters round-trip (autolock/autoungrabify/autounselectify)	gpu	6.06 ns	6.27 ns	13.1 ns	5.85 ns	27066
core: interaction tuning getters (5 calls)	gpu	7.15 ns	7.43 ns	13.7 ns	7.00 ns	22771
core: graph data() set + get (v3 restyles the graph)	v3	2.78 ms	2.82 ms	3.74 ms	2.66 ms	243
core: graph data() set + get (v3 restyles the graph)	gpu	322 ns	334 ns	403 ns	315 ns	498
core: graph scratch() set + get (v3 restyles the graph)	v3	2.79 ms	2.86 ms	3.45 ms	2.69 ms	241
core: graph scratch() set + get (v3 restyles the graph)	gpu	168 ns	170 ns	227 ns	168 ns	1002
core: batch( fn ) (empty)	v3	567 ns	688 ns	1.74 µs	562 ns	246
core: batch( fn ) (empty)	gpu	83.9 ns	86.6 ns	116 ns	82.1 ns	1957
core: startBatch + endBatch + batching	v3	532 ns	637 ns	1.58 µs	518 ns	264
core: startBatch + endBatch + batching	gpu	91.3 ns	165 ns	2.93 µs	87.4 ns	1001
core: collection()	v3	446 ns	525 ns	1.55 µs	443 ns	319
core: collection()	gpu	61.7 ns	64.5 ns	101 ns	60.4 ns	2632
core: compact() (no dead slots — the early-out)	gpu	16.9 µs	21.6 µs	97.5 µs	16.4 µs	31593
collection: element accessors (id/group/isNode/isEdge, 4 calls)	v3	60.6 ns	64.4 ns	192 ns	57.1 ns	2642
collection: element accessors (id/group/isNode/isEdge, 4 calls)	gpu	33.8 ns	36.4 ns	79.2 ns	32.6 ns	4660
collection: length + empty + nonempty (3 calls)	v3	6.60 ns	6.83 ns	11.1 ns	6.34 ns	24802
collection: length + empty + nonempty (3 calls)	gpu	1.47 ns	1.53 ns	1.94 ns	1.44 ns	106618
collection: eq/first/last (3 calls)	v3	15.1 ns	15.6 ns	25.2 ns	14.6 ns	10919
collection: eq/first/last (3 calls)	gpu	8.24 ns	8.47 ns	15.0 ns	7.93 ns	20047
collection: slice()	v3	3.80 µs	3.88 µs	4.29 µs	3.61 µs	40
collection: slice()	gpu	1.79 µs	1.81 µs	2.06 µs	1.70 µs	90
collection: toArray()	v3	8.92 µs	9.01 µs	9.25 µs	8.75 µs	15
collection: toArray()	gpu	751 ns	769 ns	884 ns	727 ns	213
collection: sort( fn )	v3	14.1 µs	14.6 µs	14.8 µs	13.8 µs	12
collection: sort( fn )	gpu	8.97 µs	8.99 µs	9.16 µs	8.85 µs	15
collection: reduce( fn )	v3	157 ns	159 ns	177 ns	119 ns	1063
collection: reduce( fn )	gpu	113 ns	115 ns	135 ns	111 ns	1500
collection: min + max ( fn ) (2 calls)	v3	11.3 µs	11.3 µs	11.6 µs	11.2 µs	11
collection: min + max ( fn ) (2 calls)	gpu	5.56 µs	5.61 µs	5.71 µs	5.48 µs	26
collection: some + every (2 calls)	v3	887 ns	904 ns	1.30 µs	862 ns	181
collection: some + every (2 calls)	gpu	565 ns	570 ns	651 ns	546 ns	297
collection: same + anySame + contains (3 calls)	v3	7.28 µs	7.30 µs	7.50 µs	7.16 µs	20
collection: same + anySame + contains (3 calls)	gpu	101 ns	103 ns	145 ns	97.2 ns	1664
collection: allAre (all match: the full walk)	v3	3.75 µs	3.75 µs	4.24 µs	3.44 µs	41
collection: allAre (all match: the full walk)	gpu	2.47 µs	2.51 µs	2.90 µs	2.41 µs	63
collection: is (none match: the full walk)	v3	5.65 µs	5.64 µs	5.88 µs	5.38 µs	26
collection: is (none match: the full walk)	gpu	2.62 µs	2.62 µs	2.94 µs	2.42 µs	61
collection: union	v3	94.7 µs	108 µs	215 µs	91.9 µs	6440
collection: union	gpu	15.1 µs	15.6 µs	16.9 µs	14.6 µs	12
collection: intersection	v3	89.7 µs	102 µs	220 µs	88.4 µs	6775
collection: intersection	gpu	3.31 µs	3.36 µs	3.79 µs	3.25 µs	46
collection: difference	v3	90.5 µs	100 µs	168 µs	88.8 µs	6863
collection: difference	gpu	3.25 µs	3.28 µs	3.82 µs	3.20 µs	47
collection: symmetricDifference	v3	24.7 µs	25.1 µs	25.5 µs	24.2 µs	12
collection: symmetricDifference	gpu	6.73 µs	6.75 µs	6.87 µs	6.58 µs	21
collection: absoluteComplement	v3	6.03 ms	6.58 ms	10.4 ms	5.84 ms	102
collection: absoluteComplement	gpu	462 µs	473 µs	724 µs	448 µs	1478
collection: byGroup	v3	979 µs	1.00 ms	1.26 ms	953 µs	697
collection: byGroup	gpu	380 µs	389 µs	605 µs	369 µs	1795
collection: diff	v3	34.5 µs	37.0 µs	80.1 µs	33.6 µs	18377
collection: diff	gpu	11.8 µs	11.8 µs	12.3 µs	10.1 µs	11
collection: indexOf	v3	32.7 ns	33.3 ns	40.8 ns	32.4 ns	5137
collection: indexOf	gpu	23.2 ns	23.9 ns	33.8 ns	22.5 ns	7138
traversal: openNeighborhood + closedNeighborhood (2 calls)	v3	9.45 µs	9.60 µs	10.9 µs	8.79 µs	14
traversal: openNeighborhood + closedNeighborhood (2 calls)	gpu	3.22 µs	3.32 µs	3.79 µs	3.13 µs	47
traversal: connectedNodes	v3	984 ns	1.07 µs	1.73 µs	957 ns	156
traversal: connectedNodes	gpu	301 ns	307 ns	407 ns	290 ns	540
traversal: sources + targets (2 calls)	v3	1.74 µs	1.89 µs	2.65 µs	1.69 µs	85
traversal: sources + targets (2 calls)	gpu	513 ns	673 ns	2.17 µs	501 ns	245
traversal: source() + target() (2 calls)	v3	71.6 ns	73.3 ns	126 ns	61.7 ns	2333
traversal: source() + target() (2 calls)	gpu	56.5 ns	58.2 ns	71.2 ns	55.0 ns	2938
traversal: edgesWith	v3	68.1 µs	69.6 µs	109 µs	66.6 µs	10008
traversal: edgesWith	gpu	56.2 µs	60.7 µs	148 µs	54.6 µs	11400
traversal: edgesTo	v3	68.6 µs	71.3 µs	125 µs	67.4 µs	9778
traversal: edgesTo	gpu	58.8 µs	59.9 µs	62.4 µs	57.4 µs	12
traversal: parallelEdges + codirectedEdges (2 calls)	v3	3.57 µs	3.86 µs	4.60 µs	3.45 µs	40
traversal: parallelEdges + codirectedEdges (2 calls)	gpu	1.07 µs	1.22 µs	1.87 µs	1.04 µs	133
traversal: roots + leaves (2 calls)	v3	5.15 ms	5.50 ms	8.79 ms	5.03 ms	121
traversal: roots + leaves (2 calls)	gpu	345 µs	350 µs	406 µs	334 µs	1996
traversal: allAreNeighbors	v3	198 µs	211 µs	280 µs	191 µs	3307
traversal: allAreNeighbors	gpu	38.9 µs	39.8 µs	42.1 µs	38.5 µs	12
traversal: isLoop + isSimple (2 calls)	v3	126 ns	128 ns	146 ns	125 ns	1332
traversal: isLoop + isSimple (2 calls)	gpu	53.9 ns	55.5 ns	69.0 ns	52.6 ns	3076
degree: indegree + outdegree (2 calls)	v3	3.50 µs	3.67 µs	4.38 µs	3.29 µs	42
degree: indegree + outdegree (2 calls)	gpu	48.8 ns	50.3 ns	75.8 ns	47.3 ns	3391
degree: minDegree + maxDegree (2 calls)	v3	15.7 ms	16.8 ms	19.9 ms	15.0 ms	37
degree: minDegree + maxDegree (2 calls)	gpu	453 µs	465 µs	711 µs	440 µs	1510
degree: minIndegree + maxOutdegree (2 calls)	v3	9.30 ms	10.1 ms	14.5 ms	8.98 ms	64
degree: minIndegree + maxOutdegree (2 calls)	gpu	434 µs	438 µs	661 µs	413 µs	1592
dimensions: renderedPosition get	v3	74.1 ns	79.3 ns	179 ns	73.3 ns	2141
dimensions: renderedPosition get	gpu	29.5 ns	33.2 ns	115 ns	28.5 ns	5080
dimensions: relativePosition get	v3	837 ns	850 ns	960 ns	826 ns	196
dimensions: relativePosition get	gpu	72.5 ns	77.9 ns	201 ns	70.1 ns	2179
dimensions: width + height + outerWidth + outerHeight (4 calls)	v3	594 ns	598 ns	653 ns	581 ns	281
dimensions: width + height + outerWidth + outerHeight (4 calls)	gpu	242 ns	246 ns	285 ns	241 ns	685
dimensions: rendered dims (4 calls)	v3	696 ns	704 ns	750 ns	692 ns	239
dimensions: rendered dims (4 calls)	gpu	237 ns	239 ns	292 ns	231 ns	711
dimensions: renderedBoundingBox	v3	1.74 µs	1.77 µs	2.26 µs	1.69 µs	90
dimensions: renderedBoundingBox	gpu	911 ns	912 ns	982 ns	892 ns	183
dimensions: midpoint + renderedMidpoint (2 calls)	gpu	1.10 µs	1.11 µs	1.22 µs	1.09 µs	146
dimensions: source/targetEndpoint (2 calls)	gpu	381 ns	384 ns	425 ns	364 ns	439
dimensions: rendered source/targetEndpoint (2 calls)	gpu	393 ns	398 ns	451 ns	377 ns	425
dimensions: shift + silentShift (round-trip)	v3	291 µs	304 µs	666 µs	283 µs	2297
dimensions: shift + silentShift (round-trip)	gpu	4.85 µs	4.87 µs	4.97 µs	4.82 µs	31
dimensions: silentPosition	v3	897 ns	910 ns	1.05 µs	874 ns	183
dimensions: silentPosition	gpu	52.7 ns	55.1 ns	81.6 ns	51.0 ns	3083
dimensions: positions( fn ) over 100	v3	169 µs	176 µs	363 µs	164 µs	3962
dimensions: positions( fn ) over 100	gpu	3.06 µs	3.07 µs	3.22 µs	2.99 µs	51
dimensions: padding + paddedWidth + paddedHeight (3 calls)	v3	1.99 µs	2.00 µs	2.13 µs	1.95 µs	80
dimensions: padding + paddedWidth + paddedHeight (3 calls)	gpu	133 ns	137 ns	175 ns	131 ns	1244
curves: isBundledBezier	gpu	20.6 ns	21.0 ns	30.3 ns	19.6 ns	8077
curves: controlPoints + renderedControlPoints (2 calls)	gpu	525 ns	535 ns	694 ns	515 ns	308
curves: segmentPoints (undefined for bezier)	gpu	33.2 ns	34.2 ns	45.4 ns	32.4 ns	5002
compound: parent + parents + ancestors (3 calls)	v3	3.03 µs	3.41 µs	4.43 µs	2.92 µs	46
compound: parent + parents + ancestors (3 calls)	gpu	835 ns	1.07 µs	2.38 µs	774 ns	152
compound: children + descendants + siblings (3 calls)	v3	8.78 µs	8.75 µs	9.37 µs	7.88 µs	16
compound: children + descendants + siblings (3 calls)	gpu	1.93 µs	2.17 µs	3.05 µs	1.87 µs	74
compound: orphans + nonorphans (2 calls)	v3	3.67 µs	3.89 µs	4.56 µs	3.54 µs	40
compound: orphans + nonorphans (2 calls)	gpu	556 ns	712 ns	2.78 µs	530 ns	229
compound: commonAncestors	v3	10.0 µs	10.2 µs	10.9 µs	9.84 µs	13
compound: commonAncestors	gpu	76.7 ns	78.9 ns	113 ns	74.5 ns	2156
compound: isParent/isChild/isChildless/isOrphan (4 calls)	v3	71.7 ns	74.8 ns	156 ns	60.6 ns	2288
compound: isParent/isChild/isChildless/isOrphan (4 calls)	gpu	45.6 ns	46.9 ns	64.3 ns	44.3 ns	3640
flags: select + unselect + selected (band)	v3	193 µs	200 µs	441 µs	188 µs	3495
flags: select + unselect + selected (band)	gpu	12.9 µs	13.0 µs	13.0 µs	12.9 µs	9
flags: selectify + unselectify + selectable	v3	195 µs	200 µs	319 µs	188 µs	3458
flags: selectify + unselectify + selectable	gpu	2.04 µs	2.05 µs	2.17 µs	2.00 µs	80
flags: lock + unlock + locked	v3	189 µs	194 µs	292 µs	185 µs	3605
flags: lock + unlock + locked	gpu	2.05 µs	2.08 µs	2.77 µs	2.02 µs	78
flags: grabify + ungrabify + grabbable	v3	190 µs	197 µs	324 µs	186 µs	3553
flags: grabify + ungrabify + grabbable	gpu	2.03 µs	2.04 µs	2.17 µs	2.01 µs	80
flags: show + hide + visible	v3	2.58 ms	2.72 ms	4.10 ms	2.50 ms	253
flags: show + hide + visible	gpu	11.5 µs	11.6 µs	11.7 µs	11.4 µs	11
flags: active/activate/unactivate (3 calls)	v3	193 µs	200 µs	398 µs	186 µs	3494
flags: active/activate/unactivate (3 calls)	gpu	12.9 µs	13.0 µs	13.0 µs	12.9 µs	9
flags: panify + unpanify + pannable	v3	192 µs	197 µs	399 µs	182 µs	3543
flags: panify + unpanify + pannable	gpu	2.02 µs	2.03 µs	2.19 µs	1.99 µs	81
flags: takesUpSpace + interactive + transparent (3 calls)	v3	196 ns	202 ns	278 ns	195 ns	849
flags: takesUpSpace + interactive + transparent (3 calls)	gpu	67.7 ns	69.3 ns	84.9 ns	66.3 ns	2468
flags: effectiveOpacity	v3	51.0 ns	52.0 ns	79.3 ns	47.7 ns	3294
flags: effectiveOpacity	gpu	23.9 ns	24.5 ns	35.9 ns	23.0 ns	6993
element data: scratch get + set	v3	1.29 µs	1.30 µs	1.41 µs	1.24 µs	128
element data: scratch get + set	gpu	176 ns	178 ns	237 ns	172 ns	961
element data: json() one element	v3	300 ns	381 ns	3.47 µs	292 ns	430
element data: json() one element	gpu	151 ns	191 ns	528 ns	147 ns	843
element data: jsons() over 100	v3	39.9 µs	40.2 µs	42.2 µs	39.3 µs	12
element data: jsons() over 100	gpu	21.3 µs	21.5 µs	24.3 µs	15.8 µs	12
element data: label()	gpu	15.4 ns	16.0 ns	30.2 ns	14.6 ns	10622
core: batchData()	v3	2.41 µs	2.55 µs	3.24 µs	2.34 µs	62
core: batchData()	gpu	281 ns	365 ns	2.30 µs	274 ns	452
core: animated()	v3	7.52 ns	7.80 ns	13.8 ns	7.22 ns	21735
core: animated()	gpu	1.53 ns	1.59 ns	2.35 ns	1.50 ns	104004
core: zoomRange get	gpu	2.16 ns	2.31 ns	6.16 ns	2.05 ns	72193
core: forceRender() (headless no-op)	v3	23.4 ns	24.6 ns	57.6 ns	22.8 ns	6949
core: forceRender() (headless no-op)	gpu	1.48 ns	1.55 ns	2.54 ns	1.45 ns	106355
core: resize() / invalidateSize()	v3	339 ns	345 ns	403 ns	329 ns	490
core: resize() / invalidateSize()	gpu	126 ns	132 ns	288 ns	121 ns	1291
core: onRender + offRender	v3	457 ns	480 ns	970 ns	437 ns	350
core: onRender + offRender	gpu	191 ns	234 ns	695 ns	188 ns	715
core: removeScratch()	v3	2.73 ms	2.77 ms	3.16 ms	2.67 ms	248
core: removeScratch()	gpu	282 ns	293 ns	559 ns	273 ns	579
core: selectionType get	v3	4.51 ns	4.71 ns	8.70 ns	4.35 ns	35812
core: selectionType get	gpu	2.26 ns	2.36 ns	4.13 ns	2.11 ns	70630
core: isReady()	v3	3.37 ns	3.46 ns	4.84 ns	3.32 ns	48536
core: isReady()	gpu	1.51 ns	1.56 ns	2.26 ns	1.48 ns	105457
core: window()	v3	4.11 ns	4.28 ns	9.14 ns	4.01 ns	39386
core: window()	gpu	1.48 ns	1.52 ns	2.01 ns	1.45 ns	108140
core: container()	v3	3.96 ns	4.13 ns	9.49 ns	3.75 ns	40859
core: container()	gpu	1.48 ns	1.54 ns	2.52 ns	1.45 ns	106489
core: destroyed()	v3	3.31 ns	3.41 ns	6.51 ns	3.18 ns	49276
core: destroyed()	gpu	1.48 ns	1.53 ns	1.95 ns	1.45 ns	107619
core: removeAllListeners()	v3	13.9 ns	14.5 ns	27.1 ns	13.4 ns	11738
core: removeAllListeners()	gpu	1.58 ns	1.64 ns	2.13 ns	1.55 ns	100166
collection: cy()	v3	7.63 ns	8.11 ns	14.8 ns	7.32 ns	21037
collection: cy()	gpu	2.16 ns	2.42 ns	6.18 ns	2.10 ns	68732
collection: indexOfId()	v3	67.4 ns	68.4 ns	91.0 ns	65.6 ns	2502
collection: indexOfId()	gpu	36.0 ns	36.7 ns	49.4 ns	35.4 ns	4658
collection: removed() + inside() (2 calls)	v3	17.2 ns	17.7 ns	27.5 ns	16.7 ns	9682
collection: removed() + inside() (2 calls)	gpu	15.4 ns	15.8 ns	18.9 ns	15.2 ns	10797
collection: silentPositions( fn ) over 100	v3	106 µs	110 µs	160 µs	103 µs	6347
collection: silentPositions( fn ) over 100	gpu	3.03 µs	3.04 µs	3.15 µs	2.98 µs	52
collection: animated()	v3	12.2 ns	12.5 ns	20.6 ns	11.7 ns	13629
collection: animated()	gpu	4.19 ns	4.35 ns	7.65 ns	4.00 ns	38813
collection: delayAnimation()	v3	450 ns	581 ns	2.06 µs	442 ns	286
collection: delayAnimation()	gpu	168 ns	173 ns	234 ns	165 ns	952
collection: removeScratch()	v3	1.80 µs	1.81 µs	1.94 µs	1.77 µs	88
collection: removeScratch()	gpu	201 ns	208 ns	393 ns	200 ns	820
collection: component()	v3	82.8 ms	85.1 ms	96.3 ms	78.5 ms	12
collection: component()	gpu	1.98 ms	2.09 ms	3.19 ms	1.91 ms	330
collection: layoutDimensions()	v3	613 ns	627 ns	847 ns	590 ns	269
collection: layoutDimensions()	gpu	170 ns	176 ns	319 ns	165 ns	968
degree: maxIndegree + minOutdegree (2 calls)	v3	9.27 ms	10.1 ms	14.2 ms	8.94 ms	66
degree: maxIndegree + minOutdegree (2 calls)	gpu	428 µs	447 µs	698 µs	414 µs	1557
flags: grabbed + hidden + inactive (3 calls)	v3	86.5 ns	87.1 ns	100 ns	82.3 ns	1962
flags: grabbed + hidden + inactive (3 calls)	gpu	42.7 ns	44.3 ns	86.8 ns	41.0 ns	3860
curves: renderedSegmentPoints	gpu	36.1 ns	37.0 ns	55.1 ns	34.8 ns	4629
layout contract: edgeSlots + endpoints + degreeOf	gpu	34.7 µs	36.7 µs	64.1 µs	29.8 µs	18758
formats: isColumnarElements + isSerializedElements + isPackedIds	gpu	4.61 ns	4.75 ns	8.20 ns	4.60 ns	35671

compaction N=20,000 · 18.5 s

cpu pick: background miss (full highWater walk) — peak

221 µs

cpu pick: background miss (full highWater walk) — compacted

51.0 µs

cpu pick: node hit — peak

226 µs

cpu pick: node hit — compacted

49.7 µs

isCurrent: current ref (forwards present vs absent) — no forwards (peak)

5.92 ns

isCurrent: current ref (forwards present vs absent) — forwards present (compacted)

5.95 ns

isCurrent: stale-ref repair (chase + rewrite, alloc included) — current (baseline alloc)

8.80 ns

isCurrent: stale-ref repair (chase + rewrite, alloc included) — stale → repair

49.3 ns

control: nodes({ selected }) scan (order-list, expected parity) — peak

24.8 µs

control: nodes({ selected }) scan (order-list, expected parity) — compacted

21.3 µs

control: whole-graph boundingBox (expected parity) — peak

96.3 µs

control: whole-graph boundingBox (expected parity) — compacted

95.5 µs

1 ns10 ns100 ns1 µs10 µs100 µs1 ms

All stats (table view)

group	bench	p50	avg	p99	min	samples
cpu pick: background miss (full highWater walk)	peak	221 µs	225 µs	269 µs	221 µs	3125
cpu pick: background miss (full highWater walk)	compacted	51.0 µs	51.2 µs	51.7 µs	49.9 µs	12
cpu pick: node hit	peak	226 µs	230 µs	292 µs	221 µs	3058
cpu pick: node hit	compacted	49.7 µs	50.0 µs	50.3 µs	49.2 µs	12
isCurrent: current ref (forwards present vs absent)	no forwards (peak)	5.92 ns	6.10 ns	10.0 ns	5.74 ns	27793
isCurrent: current ref (forwards present vs absent)	forwards present (compacted)	5.95 ns	6.11 ns	9.72 ns	5.74 ns	27744
isCurrent: stale-ref repair (chase + rewrite, alloc included)	current (baseline alloc)	8.80 ns	9.43 ns	21.7 ns	8.38 ns	17900
isCurrent: stale-ref repair (chase + rewrite, alloc included)	stale → repair	49.3 ns	51.0 ns	72.6 ns	48.8 ns	3352
control: nodes({ selected }) scan (order-list, expected parity)	peak	24.8 µs	26.3 µs	31.1 µs	21.4 µs	12
control: nodes({ selected }) scan (order-list, expected parity)	compacted	21.3 µs	22.0 µs	24.5 µs	20.6 µs	12
control: whole-graph boundingBox (expected parity)	peak	96.3 µs	102 µs	198 µs	93.2 µs	6830
control: whole-graph boundingBox (expected parity)	compacted	95.5 µs	97.4 µs	140 µs	92.6 µs	7166

compound N=2,000 · 9.3 sv4 (gpu) v3

parent drag (shift subtree + bb settle)

186×

child drag (move + parent re-derive)

51×

reparent round-trip

70×

1 µs10 µs100 µs1 ms

All stats (table view)

group	bench	p50	avg	p99	min	samples
parent drag (shift subtree + bb settle)	v3	253 µs	285 µs	672 µs	238 µs	2442
parent drag (shift subtree + bb settle)	gpu	1.36 µs	1.41 µs	2.16 µs	1.32 µs	114
child drag (move + parent re-derive)	v3	80.1 µs	83.5 µs	127 µs	75.3 µs	8216
child drag (move + parent re-derive)	gpu	1.57 µs	1.59 µs	1.72 µs	1.54 µs	102
reparent round-trip	v3	82.2 µs	88.9 µs	155 µs	79.8 µs	7744
reparent round-trip	gpu	1.18 µs	1.20 µs	1.58 µs	1.13 µs	136

curves N=2,000 · 0.6 s

curve premium: build (style apply + derivation) — curved

25.7 ms

curve premium: build (style apply + derivation) — straight

22.6 ms

curve premium: node.position x 1000 (drag) — curved

589 µs

curve premium: node.position x 1000 (drag) — straight

253 µs

curve premium: nodes().positions() (bulk move) — curved

304 µs

curve premium: nodes().positions() (bulk move) — straight

297 µs

curve premium: first read after a bulk move (cold) — curved

7.56 ms

curve premium: first read after a bulk move (cold) — straight

2.99 ms

curve premium: same read again (warm) — curved

2.59 ms

curve premium: same read again (warm) — straight

1.64 ms

curve premium: edge.midpoint x 4000 — curved

1.89 ms

curve premium: edge.midpoint x 4000 — straight

3.01 ms

curve premium: edge.controlPoints x 4000 — curved

1.89 ms

curve premium: edge.controlPoints x 4000 — straight

1.11 ms

curve premium: cy.elements().boundingBox() x 20 — curved

35.1 ms

curve premium: cy.elements().boundingBox() x 20 — straight

21.9 ms

curve premium: cy.fit() x 20 (store scan) — curved

3.94 ms

curve premium: cy.fit() x 20 (store scan) — straight

3.58 ms

curve premium: elementsInBox x 20 (half the graph) — curved

23.7 ms

curve premium: elementsInBox x 20 (half the graph) — straight

5.86 ms

curve premium: edge.hide/show x 1000 (bundle re-fan) — curved

3.98 ms

curve premium: edge.hide/show x 1000 (bundle re-fan) — straight

2.14 ms

100 µs1 ms10 ms100 ms

All stats (table view)

group	bench	p50	avg	p99	min	samples
curve premium: build (style apply + derivation)	curved	25.7 ms	25.7 ms	25.7 ms	25.7 ms	1
curve premium: build (style apply + derivation)	straight	22.6 ms	22.6 ms	22.6 ms	22.6 ms	1
curve premium: node.position x 1000 (drag)	curved	589 µs	589 µs	589 µs	589 µs	1
curve premium: node.position x 1000 (drag)	straight	253 µs	253 µs	253 µs	253 µs	1
curve premium: nodes().positions() (bulk move)	curved	304 µs	304 µs	304 µs	304 µs	1
curve premium: nodes().positions() (bulk move)	straight	297 µs	297 µs	297 µs	297 µs	1
curve premium: first read after a bulk move (cold)	curved	7.56 ms	7.56 ms	7.56 ms	7.56 ms	1
curve premium: first read after a bulk move (cold)	straight	2.99 ms	2.99 ms	2.99 ms	2.99 ms	1
curve premium: same read again (warm)	curved	2.59 ms	2.59 ms	2.59 ms	2.59 ms	1
curve premium: same read again (warm)	straight	1.64 ms	1.64 ms	1.64 ms	1.64 ms	1
curve premium: edge.midpoint x 4000	curved	1.89 ms	1.89 ms	1.89 ms	1.89 ms	1
curve premium: edge.midpoint x 4000	straight	3.01 ms	3.01 ms	3.01 ms	3.01 ms	1
curve premium: edge.controlPoints x 4000	curved	1.89 ms	1.89 ms	1.89 ms	1.89 ms	1
curve premium: edge.controlPoints x 4000	straight	1.11 ms	1.11 ms	1.11 ms	1.11 ms	1
curve premium: cy.elements().boundingBox() x 20	curved	35.1 ms	35.1 ms	35.1 ms	35.1 ms	1
curve premium: cy.elements().boundingBox() x 20	straight	21.9 ms	21.9 ms	21.9 ms	21.9 ms	1
curve premium: cy.fit() x 20 (store scan)	curved	3.94 ms	3.94 ms	3.94 ms	3.94 ms	1
curve premium: cy.fit() x 20 (store scan)	straight	3.58 ms	3.58 ms	3.58 ms	3.58 ms	1
curve premium: elementsInBox x 20 (half the graph)	curved	23.7 ms	23.7 ms	23.7 ms	23.7 ms	1
curve premium: elementsInBox x 20 (half the graph)	straight	5.86 ms	5.86 ms	5.86 ms	5.86 ms	1
curve premium: edge.hide/show x 1000 (bundle re-fan)	curved	3.98 ms	3.98 ms	3.98 ms	3.98 ms	1
curve premium: edge.hide/show x 1000 (bundle re-fan)	straight	2.14 ms	2.14 ms	2.14 ms	2.14 ms	1

arrows N=2,000 · 3.3 s

arrow premium: build (style apply + arrow fold) — arrowed

26.7 ms

arrow premium: build (style apply + arrow fold) — none

21.0 ms

arrow premium: cy.style() re-apply x 4 — arrowed

40.8 ms

arrow premium: cy.style() re-apply x 4 — none

39.6 ms

arrow premium: node.position x 200000 (drag) — arrowed

24.5 ms

arrow premium: node.position x 200000 (drag) — none

24.9 ms

arrow premium: edge.sourceEndpoint x 200000 — arrowed

61.7 ms

arrow premium: edge.sourceEndpoint x 200000 — none

59.8 ms

arrow premium: edge.targetEndpoint x 200000 — arrowed

60.8 ms

arrow premium: edge.targetEndpoint x 200000 — none

60.3 ms

arrow premium: edge.midpoint x 200000 — arrowed

151 ms

arrow premium: edge.midpoint x 200000 — none

155 ms

arrow premium: edge.style('target-arrow-shape') x 200000 — arrowed

35.7 ms

arrow premium: edge.style('target-arrow-shape') x 200000 — none

30.9 ms

arrow premium: cy.elements().boundingBox() x 20 — arrowed

21.6 ms

arrow premium: cy.elements().boundingBox() x 20 — none

21.7 ms

arrow premium: elementsInBox x 20 (half the graph) — arrowed

5.61 ms

arrow premium: elementsInBox x 20 (half the graph) — none

5.58 ms

arrow premium: control: sourceEndpoint, bezier vs straight — arrowed

76.0 ms

arrow premium: control: sourceEndpoint, bezier vs straight — none

60.6 ms

arrow premium: edge column bytes — B/edge

156 ms

1 ms10 ms100 ms1 s

All stats (table view)

group	bench	p50	avg	p99	min	samples
arrow premium: build (style apply + arrow fold)	arrowed	26.7 ms	26.7 ms	26.7 ms	26.7 ms	1
arrow premium: build (style apply + arrow fold)	none	21.0 ms	21.0 ms	21.0 ms	21.0 ms	1
arrow premium: cy.style() re-apply x 4	arrowed	40.8 ms	40.8 ms	40.8 ms	40.8 ms	1
arrow premium: cy.style() re-apply x 4	none	39.6 ms	39.6 ms	39.6 ms	39.6 ms	1
arrow premium: node.position x 200000 (drag)	arrowed	24.5 ms	24.5 ms	24.5 ms	24.5 ms	1
arrow premium: node.position x 200000 (drag)	none	24.9 ms	24.9 ms	24.9 ms	24.9 ms	1
arrow premium: edge.sourceEndpoint x 200000	arrowed	61.7 ms	61.7 ms	61.7 ms	61.7 ms	1
arrow premium: edge.sourceEndpoint x 200000	none	59.8 ms	59.8 ms	59.8 ms	59.8 ms	1
arrow premium: edge.targetEndpoint x 200000	arrowed	60.8 ms	60.8 ms	60.8 ms	60.8 ms	1
arrow premium: edge.targetEndpoint x 200000	none	60.3 ms	60.3 ms	60.3 ms	60.3 ms	1
arrow premium: edge.midpoint x 200000	arrowed	151 ms	151 ms	151 ms	151 ms	1
arrow premium: edge.midpoint x 200000	none	155 ms	155 ms	155 ms	155 ms	1
arrow premium: edge.style('target-arrow-shape') x 200000	arrowed	35.7 ms	35.7 ms	35.7 ms	35.7 ms	1
arrow premium: edge.style('target-arrow-shape') x 200000	none	30.9 ms	30.9 ms	30.9 ms	30.9 ms	1
arrow premium: cy.elements().boundingBox() x 20	arrowed	21.6 ms	21.6 ms	21.6 ms	21.6 ms	1
arrow premium: cy.elements().boundingBox() x 20	none	21.7 ms	21.7 ms	21.7 ms	21.7 ms	1
arrow premium: elementsInBox x 20 (half the graph)	arrowed	5.61 ms	5.61 ms	5.61 ms	5.61 ms	1
arrow premium: elementsInBox x 20 (half the graph)	none	5.58 ms	5.58 ms	5.58 ms	5.58 ms	1
arrow premium: control: sourceEndpoint, bezier vs straight	arrowed	76.0 ms	76.0 ms	76.0 ms	76.0 ms	1
arrow premium: control: sourceEndpoint, bezier vs straight	none	60.6 ms	60.6 ms	60.6 ms	60.6 ms	1
arrow premium: edge column bytes	B/edge	156 ms	156 ms	156 ms	156 ms	1

labels N=20,000 · 0.5 s

labels: breakLines x 20000 — gpu

78.8 ms

labels: estimateBlock x 20000 — gpu

89.0 ms

labels: addNode x 20000 — gpu

29.2 ms

labels: setLabel (build) x 20000 — gpu

110 ms

labels: setLabel (rewrite) x 20000 — gpu

101 ms

labels: boundingBox with 20000 label terms — gpu

3.20 ms

1 ms10 ms100 ms1 s

All stats (table view)

group	bench	p50	avg	p99	min	samples
labels: breakLines x 20000	gpu	78.8 ms	78.8 ms	78.8 ms	78.8 ms	1
labels: estimateBlock x 20000	gpu	89.0 ms	89.0 ms	89.0 ms	89.0 ms	1
labels: addNode x 20000	gpu	29.2 ms	29.2 ms	29.2 ms	29.2 ms	1
labels: setLabel (build) x 20000	gpu	110 ms	110 ms	110 ms	110 ms	1
labels: setLabel (rewrite) x 20000	gpu	101 ms	101 ms	101 ms	101 ms	1
labels: boundingBox with 20000 label terms	gpu	3.20 ms	3.20 ms	3.20 ms	3.20 ms	1

transitions N=2,000 · 9.5 s

transition: auto-extent shift, whole-channel re-derive (2000 slots) — transitions off

2.21 ms

transition: auto-extent shift, whole-channel re-derive (2000 slots) — transitions on

4.54 ms

transition: bulk tween tick (2000 slots, one manager tick) — tick

127 µs

transition: explicit-domain write (O(changed) — one element) — transitions off

2.35 µs

transition: explicit-domain write (O(changed) — one element) — transitions on

6.15 µs

transition: sheet swap, whole-graph apply (2000 slots diff + tween) — transitions off

9.97 ms

transition: sheet swap, whole-graph apply (2000 slots diff + tween) — transitions on

13.0 ms

1 µs10 µs100 µs1 ms10 ms100 ms

All stats (table view)

group	bench	p50	avg	p99	min	samples
transition: auto-extent shift, whole-channel re-derive (2000 slots)	transitions off	2.21 ms	2.31 ms	4.36 ms	2.15 ms	297
transition: auto-extent shift, whole-channel re-derive (2000 slots)	transitions on	4.54 ms	4.84 ms	7.87 ms	4.24 ms	138
transition: bulk tween tick (2000 slots, one manager tick)	tick	127 µs	131 µs	207 µs	123 µs	5329
transition: explicit-domain write (O(changed) — one element)	transitions off	2.35 µs	2.43 µs	3.41 µs	2.29 µs	65
transition: explicit-domain write (O(changed) — one element)	transitions on	6.15 µs	6.06 µs	6.85 µs	4.96 µs	24
transition: sheet swap, whole-graph apply (2000 slots diff + tween)	transitions off	9.97 ms	10.0 ms	11.4 ms	9.64 ms	65
transition: sheet swap, whole-graph apply (2000 slots diff + tween)	transitions on	13.0 ms	12.9 ms	15.7 ms	11.3 ms	50

geometry-tween N=2,000 · 11.1 s

size tick: unlabelled (2000 slots) — tick

589 µs

size tick: labelled, center/center (re-anchor early-out) (2000 slots) — tick

636 µs

size tick: labelled, hanging (sidecar rewrite per tick) (2000 slots) — tick

1.26 ms

edge width tick (4000 slots) — bare

297 µs

edge width tick (4000 slots) — with rides (casing + match-line arrows)

706 µs

padding tick + auto-bounds flush (250 parents × 8 children) — tick + flush

179 µs

font-size tick (2000 labels) — wrap none (dims scale-patch)

359 µs

font-size tick (2000 labels) — wrapped (re-break per tick)

5.69 ms

baseline: paint (opacity) tick (2000 slots) — tick

142 µs

100 µs1 ms10 ms

All stats (table view)

group	bench	p50	avg	p99	min	samples
size tick: unlabelled (2000 slots)	tick	589 µs	611 µs	930 µs	582 µs	1121
size tick: labelled, center/center (re-anchor early-out) (2000 slots)	tick	636 µs	675 µs	1.25 ms	617 µs	1037
size tick: labelled, hanging (sidecar rewrite per tick) (2000 slots)	tick	1.26 ms	1.34 ms	2.34 ms	1.13 ms	502
edge width tick (4000 slots)	bare	297 µs	306 µs	442 µs	293 µs	2276
edge width tick (4000 slots)	with rides (casing + match-line arrows)	706 µs	718 µs	863 µs	686 µs	971
padding tick + auto-bounds flush (250 parents × 8 children)	tick + flush	179 µs	187 µs	383 µs	173 µs	3738
font-size tick (2000 labels)	wrap none (dims scale-patch)	359 µs	400 µs	921 µs	341 µs	1728
font-size tick (2000 labels)	wrapped (re-break per tick)	5.69 ms	5.77 ms	6.41 ms	5.46 ms	116
baseline: paint (opacity) tick (2000 slots)	tick	142 µs	148 µs	261 µs	138 µs	4726