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	1f324088 fix(65.9b): style-bundle.mjs has been a module-level SyntaxError since round 42 — every all-profile run published it as a failure
Machine id	5cf3f79c

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

sweep: edges() @ 2k

304501×

sweep: nodes() @ 2k

167819×

core: edges() @ 2k

93617×

core: nodes() @ 2k

51120×

sweep: elements() @ 2k

29945×

core: elements() @ 2k

23406×

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

17043×

core: removeScratch() @ 2k

9536×

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

8273×

compare: same() @ 2k

7609×

trav: successors() @ 2k

757×

trav: predecessors() @ 2k

736×

mut-bulk: hide + show @ 2k

283×

flags: show + hide + visible @ 2k

208×

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

186×

degree: nodes.totalDegree() @ 2k

167×

154×

degree: nodes.maxDegree() @ 2k

141×

compound: commonAncestors @ 2k

128×

degree: node.degree() @ 2k

119×

viewport: reset() @ 2k

109×

mut-bulk: positions(obj) @ 2k

105×

mut-bulk: shift @ 2k

100×

mut-bulk: positions(fn) @ 2k

99×

flags: panify + unpanify + pannable @ 2k

97×

flags: grabify + ungrabify + grabbable @ 2k

96×

flags: selectify + unselectify + selectable @ 2k

96×

flags: lock + unlock + locked @ 2k

94×

core: options() @ 2k

84×

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

75×

algo: closenessCentrality (one root) @ 500

75×

reparent round-trip @ 2k

74×

algo: closenessCentrality (one root) @ 2k

74×

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

70×

algo: bfs (whole graph) @ 500

65×

algo: dfs (whole graph) @ 500

63×

algo: bfs (whole graph) @ 2k

62×

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

62×

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

61×

algo: dfs (whole graph) @ 2k

61×

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

59×

57×

algo: closenessCentrality (weighted, one root) @ 500

57×

dimensions: positions( fn ) over 100 @ 2k

57×

algo: dijkstra + pathTo @ 2k

57×

algo: dijkstra + pathTo @ 500

55×

mut-bulk: lock + unlock @ 2k

52×

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

51×

collection: component() @ 2k

42×

traverse: components() @ 2k

40×

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

40×

algo: degreeCentralityNormalized @ 500

40×

algo: tarjan SCC @ 500

38×

algo: tarjan SCC @ 2k

38×

algo: degreeCentralityNormalized @ 2k

38×

37×

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

35×

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

34×

collection: silentPositions( fn ) over 100 @ 2k

34×

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

33×

algo: degreeCentrality (one root) @ 500

32×

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

32×

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

31×

algo: degreeCentrality (one root) @ 2k

30×

set: difference() @ 2k

29×

collection: difference @ 2k

28×

collection: intersection @ 2k

28×

algo: kruskal @ 2k

27×

set: intersection() @ 2k

27×

algo: hopcroft-tarjan biconnected @ 500

26×

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

26×

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

26×

algo: hopcroft-tarjan biconnected @ 2k

26×

26×

algo: kruskal @ 500

25×

emit: position write, no listeners @ 2k

24×

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

24×

mut: position set @ 2k

24×

trav: band roots() @ 2k

22×

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

21×

algo: bellmanFord @ 500

21×

algo: bellmanFord @ 2k

20×

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

19×

data: bulk write, numeric column @ 2k

19×

viewport: center( eles ) @ 2k

19×

18×

core: add + remove @ 2k

17×

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

17×

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

16×

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

16×

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

16×

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

16×

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

16×

dimensions: silentPosition @ 2k

15×

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

15×

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

15×

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

15×

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

14×

style: renderedStyle (one prop) @ 2k

14×

mut-bulk: data set @ 2k

14×

collection: absoluteComplement @ 2k

13×

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

13×

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

13×

dimensions: padding + paddedWidth + paddedHeight (3 calls) @ 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×

dimensions: relativePosition get @ 2k

11×

algo: betweennessCentrality (unweighted) @ 2k

11×

algo: betweennessCentrality (unweighted) @ 500

11×

11×

collection: removeScratch() @ 2k

9.6×

9.5×

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

9.4×

algo: betweennessCentrality (weighted) @ 500

9.0×

core: removeAllListeners() @ 2k

8.7×

algo: kMedoids @ 500

8.7×

core: batchData() @ 2k

8.6×

algo: fuzzyCMeans @ 500

8.6×

query: predicate function (both sides) @ 2k

8.5×

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

8.2×

emit: 1 core listener @ 2k

8.1×

element data: scratch get + set @ 2k

7.8×

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

7.7×

core: collection() @ 2k

7.3×

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

7.3×

data: removeData (whole collection) @ 2k

7.2×

core: collection() @ 2k

7.1×

mut-bulk: select + unselect @ 2k

7.1×

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

7.0×

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

6.7×

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

6.5×

collection: union @ 2k

6.3×

emit: 10 core listeners @ 2k

6.3×

style: whole-object style() @ 2k

6.2×

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

6.2×

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

6.2×

bounds: whole-graph boundingBox() @ 2k

6.0×

core: startBatch + endBatch + batching @ 2k

5.8×

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

5.6×

algo: hierholzer (directed) @ 500

5.4×

traverse: incomers() @ 2k

5.3×

trav: outgoers() @ 2k

5.3×

algo: affinityPropagation @ 500

5.2×

traverse: outgoers() @ 2k

5.2×

set: union() @ 2k

5.2×

trav: incomers() @ 2k

5.1×

traversal: allAreNeighbors @ 2k

5.0×

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

4.9×

core: animated() @ 2k

4.9×

trav: band neighborhood() @ 2k

4.9×

compare: contains() @ 2k

4.7×

viewport: zoom() get @ 2k

4.6×

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

4.6×

style: bypass set + removeStyle round-trip @ 2k

4.3×

mut: data set @ 2k

4.2×

trav: neighborhood() @ 2k

4.0×

algo: kMeans (fixed centroids) @ 500

3.9×

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

3.9×

traverse: neighborhood() @ 2k

3.8×

mut: select + unselect @ 2k

3.8×

collection: symmetricDifference @ 2k

3.8×

algo: aStar @ 500

3.6×

collection: cy() @ 2k

3.6×

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

3.5×

algo: aStar @ 2k

3.4×

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

3.3×

traversal: connectedNodes @ 2k

3.3×

collection: diff @ 2k

3.2×

style: bypass set (ele.style name,value) @ 2k

3.2×

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

3.1×

collection: layoutDimensions() @ 2k

3.1×

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

3.0×

algo: hierholzer (directed) @ 2k

3.0×

collection: animated() @ 2k

2.9×

dimensions: rendered dims (4 calls) @ 2k

2.9×

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

2.8×

trav: band connectedNodes() @ 2k

2.8×

anim: viewport animate + stop @ 2k

2.8×

core: window() @ 2k

2.8×

collection: delayAnimation() @ 2k

2.8×

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

2.7×

load: cy.json() @ 2k

2.7×

collection: byGroup @ 2k

2.7×

trav: connectedEdges() @ 2k

2.6×

core: container() @ 2k

2.6×

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

2.6×

traverse: connectedEdges() @ 2k

2.6×

flags: effectiveOpacity @ 2k

2.6×

trav: closedNeighborhood() @ 2k

2.5×

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

2.5×

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

2.5×

algo: hierarchicalClustering (threshold) @ 500

2.5×

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

2.5×

iter: map() @ 2k

2.4×

core: onRender + offRender @ 2k

2.3×

data: read the whole object @ 2k

2.3×

dimensions: renderedPosition get @ 2k

2.2×

collection: slice() @ 2k

2.2×

core: destroyed() @ 2k

2.2×

element data: json() one element @ 2k

2.2×

core: isReady() @ 2k

2.2×

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

2.2×

style: read one prop (width) @ 2k

2.1×

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

2.1×

style: numericStyle @ 2k

2.00×

core: selectionType get @ 2k

1.99×

viewport: zoom( n ) set @ 2k

1.99×

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

1.96×

json: element @ 2k

1.93×

algo: pageRank (20 iters) @ 500

1.92×

iter: filter(fn) @ 2k

1.87×

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

1.87×

dimensions: renderedBoundingBox @ 2k

1.86×

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

1.85×

viewport: pan( p ) set @ 2k

1.84×

core: zoom set @ 2k

1.79×

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

1.78×

element data: jsons() over 100 @ 2k

1.75×

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

1.75×

viewport: panBy() @ 2k

1.70×

style( 'text-max-width' ) (was #74 in the switch) @ 2k

1.69×

sweep: filter(fn) @ 2k

1.68×

algo: closenessCentralityNormalized @ 500

1.68×

style( 'text-wrap' ) (was #73 in the switch) @ 2k

1.67×

viewport: renderedExtent() @ 2k

1.65×

collection: indexOfId() @ 2k

1.63×

bounds: one node boundingBox() @ 2k

1.62×

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

1.61×

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

1.61×

algo: floydWarshall @ 500

1.60×

trav: band sources() @ 2k

1.60×

collection: sort( fn ) @ 2k

1.60×

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

1.56×

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

1.56×

viewport: extent() @ 2k

1.52×

position: get @ 2k

1.51×

style: read one prop (color) @ 2k

1.48×

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

1.47×

anim: start + stop, one element @ 2k

1.46×

trav: band connectedEdges() @ 2k

1.42×

collection: reduce( fn ) @ 2k

1.40×

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

1.36×

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

1.36×

data: read one numeric key @ 2k

1.35×

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

1.34×

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

1.33×

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

1.31×

collection: indexOf @ 2k

1.29×

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

1.27×

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

1.27×

traversal: edgesWith @ 2k

1.25×

traversal: edgesTo @ 2k

1.24×

style( 'target-distance-from-node' ) (was #150 in the switch) @ 2k

1.22×

style( 'taxi-radius' ) (was #142 in the switch) @ 2k

1.22×

style( 'border-width' ) (was #6 in the switch) @ 2k

1.19×

core: filter(fn) @ 2k

1.18×

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

1.17×

data: get @ 2k

1.16×

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

1.14×

core: getElementById() @ 2k

1.14×

style( 'background-color' ) (was #4 in the switch) @ 2k

1.14×

core: gating getters (5 calls) @ 2k

1.14×

algo: markovClustering @ 500

1.10×

core: extent() @ 2k

1.10×

viewport: pan() get @ 2k

1.08×

iter: forEach() @ 2k

1.04×

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)	65×	62×
algo: dfs (whole graph)	63×	61×
algo: dijkstra + pathTo	55×	57×
algo: aStar	3.6×	3.4×
algo: bellmanFord	21×	20×
algo: kruskal	25×	27×
algo: tarjan SCC	38×	38×
algo: hopcroft-tarjan biconnected	26×	26×
algo: hierholzer (directed)	5.4×	3.0×
algo: betweennessCentrality (unweighted)	11×	11×
algo: degreeCentralityNormalized	40×	38×
algo: closenessCentrality (one root)	75×	74×
algo: degreeCentrality (one root)	32×	30×
algo: pageRank (20 iters)	1.92×	–
algo: closenessCentrality (weighted, one root)	57×	–
algo: floydWarshall	1.60×	–
algo: closenessCentralityNormalized	1.68×	–
algo: markovClustering	1.10×	–
algo: hierarchicalClustering (threshold)	2.5×	–
algo: kMeans (fixed centroids)	3.9×	–
algo: kMedoids	8.7×	–
algo: fuzzyCMeans	8.6×	–
algo: affinityPropagation	5.2×	–
algo: kargerStein	1.03×	–
algo: betweennessCentrality (weighted)	9.0×	–

core+collection N=2,000 · 58.4 sv4 (gpu) v3

core: getElementById()

1.14×

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

26×

core: nodes()

51120×

core: edges()

93617×

core: elements()

23406×

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

5.6×

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

14×

core: filter(fn)

1.18×

core: collection()

7.1×

core: pan() get (x32)

1.34×

core: extent()

1.10×

core: add + remove

17×

core: zoom set

1.79×

traverse: neighborhood()

3.8×

traverse: connectedEdges()

2.6×

traverse: outgoers()

5.2×

traverse: incomers()

5.3×

traverse: components()

40×

degree: node.degree()

119×

degree: nodes.totalDegree()

167×

degree: nodes.maxDegree()

141×

data: get

1.16×

position: get

1.51×

json: element

1.93×

set: union()

5.2×

set: intersection()

27×

set: difference()

29×

iter: map()

2.4×

iter: forEach()

1.04×

iter: filter(fn)

1.87×

compare: same()

7609×

compare: contains()

4.7×

mut: data set

4.2×

mut: position set

24×

mut: select + unselect

3.8×

1 ns10 ns100 ns1 µs10 µs100 µs1 ms10 ms100 ms

All stats (table view)

group	bench	p50	avg	p99	min	samples
core: getElementById()	v3	95.0 ns	96.8 ns	117 ns	91.5 ns	1772
core: getElementById()	gpu	83.1 ns	86.1 ns	126 ns	80.8 ns	1990
core: id lookup ($("#id") vs $id())	v3	2.03 µs	2.37 µs	3.57 µs	1.93 µs	65
core: id lookup ($("#id") vs $id())	gpu	78.4 ns	84.1 ns	161 ns	76.3 ns	2032
core: nodes()	v3	419 µs	439 µs	806 µs	413 µs	1591
core: nodes()	gpu	8.20 ns	8.76 ns	17.0 ns	7.88 ns	19358
core: edges()	v3	742 µs	755 µs	934 µs	722 µs	925
core: edges()	gpu	7.93 ns	8.28 ns	15.9 ns	7.57 ns	20572
core: elements()	v3	162 µs	169 µs	284 µs	160 µs	4130
core: elements()	gpu	6.94 ns	7.28 ns	23.9 ns	6.63 ns	23274
core: nodes ($("node") vs filter({group}))	v3	562 µs	589 µs	1.25 ms	548 µs	1185
core: nodes ($("node") vs filter({group}))	gpu	100 µs	106 µs	262 µs	71.0 µs	6439
core: selected ($(":selected") vs filter({selected}))	v3	364 µs	380 µs	697 µs	345 µs	1830
core: selected ($(":selected") vs filter({selected}))	gpu	25.6 µs	26.3 µs	26.5 µs	25.3 µs	12
core: filter(fn)	v3	415 µs	429 µs	640 µs	408 µs	1629
core: filter(fn)	gpu	351 µs	384 µs	857 µs	343 µs	1796
core: collection()	v3	476 ns	576 ns	1.71 µs	448 ns	290
core: collection()	gpu	67.4 ns	73.5 ns	288 ns	65.6 ns	2298
core: pan() get (x32)	v3	45.6 ns	45.7 ns	48.4 ns	43.4 ns	3617
core: pan() get (x32)	gpu	34.1 ns	36.4 ns	55.2 ns	33.1 ns	4698
core: extent()	v3	22.8 ns	39.1 ns	144 ns	21.8 ns	4271
core: extent()	gpu	20.8 ns	28.0 ns	111 ns	19.5 ns	5922
core: add + remove	v3	130 µs	137 µs	278 µs	121 µs	5034
core: add + remove	gpu	7.56 µs	8.57 µs	22.5 µs	7.02 µs	73631
core: zoom set	v3	1.26 µs	1.28 µs	1.46 µs	1.20 µs	128
core: zoom set	gpu	704 ns	712 ns	931 ns	645 ns	228
traverse: neighborhood()	v3	3.60 µs	3.82 µs	4.62 µs	3.39 µs	40
traverse: neighborhood()	gpu	942 ns	1.01 µs	1.66 µs	906 ns	156
traverse: connectedEdges()	v3	1.01 µs	1.11 µs	1.90 µs	974 ns	147
traverse: connectedEdges()	gpu	388 ns	410 ns	1.24 µs	378 ns	405
traverse: outgoers()	v3	2.35 µs	2.54 µs	3.40 µs	2.22 µs	63
traverse: outgoers()	gpu	453 ns	540 ns	2.06 µs	441 ns	300
traverse: incomers()	v3	2.44 µs	2.62 µs	3.35 µs	2.27 µs	61
traverse: incomers()	gpu	462 ns	582 ns	2.15 µs	453 ns	278
traverse: components()	v3	82.5 ms	85.8 ms	95.1 ms	80.0 ms	12
traverse: components()	gpu	2.04 ms	2.17 ms	3.61 ms	2.00 ms	307
degree: node.degree()	v3	3.37 µs	3.65 µs	4.37 µs	3.24 µs	42
degree: node.degree()	gpu	28.3 ns	29.5 ns	40.0 ns	27.2 ns	5788
degree: nodes.totalDegree()	v3	7.29 ms	8.16 ms	11.6 ms	6.95 ms	82
degree: nodes.totalDegree()	gpu	43.6 µs	44.8 µs	79.7 µs	42.4 µs	15391
degree: nodes.maxDegree()	v3	7.26 ms	8.44 ms	13.2 ms	6.92 ms	79
degree: nodes.maxDegree()	gpu	51.6 µs	52.9 µs	88.2 µs	50.6 µs	12735
data: get	v3	30.0 ns	30.9 ns	48.7 ns	28.6 ns	5535
data: get	gpu	26.0 ns	26.7 ns	36.9 ns	24.9 ns	6411
position: get	v3	31.4 ns	33.5 ns	126 ns	31.0 ns	5110
position: get	gpu	20.8 ns	25.2 ns	71.1 ns	20.1 ns	6540
json: element	v3	294 ns	392 ns	3.62 µs	287 ns	423
json: element	gpu	152 ns	160 ns	338 ns	148 ns	1018
set: union()	v3	94.2 µs	110 µs	223 µs	92.5 µs	6115
set: union()	gpu	18.3 µs	19.1 µs	19.2 µs	17.4 µs	12
set: intersection()	v3	91.4 µs	103 µs	157 µs	89.6 µs	6596
set: intersection()	gpu	3.41 µs	3.45 µs	3.69 µs	3.38 µs	43
set: difference()	v3	99.0 µs	114 µs	176 µs	95.0 µs	6021
set: difference()	gpu	3.41 µs	3.45 µs	3.90 µs	3.34 µs	45
iter: map()	v3	23.9 µs	25.3 µs	26.2 µs	23.7 µs	12
iter: map()	gpu	10.1 µs	10.2 µs	10.2 µs	10.1 µs	13
iter: forEach()	v3	3.44 µs	3.78 µs	4.74 µs	3.43 µs	36
iter: forEach()	gpu	3.31 µs	3.34 µs	3.60 µs	3.27 µs	45
iter: filter(fn)	v3	240 µs	248 µs	338 µs	237 µs	2823
iter: filter(fn)	gpu	128 µs	133 µs	212 µs	125 µs	5273
compare: same()	v3	113 µs	116 µs	171 µs	108 µs	6035
compare: same()	gpu	14.8 ns	16.0 ns	30.2 ns	14.3 ns	10690
compare: contains()	v3	6.07 µs	6.16 µs	6.44 µs	5.98 µs	24
compare: contains()	gpu	1.29 µs	1.30 µs	1.40 µs	1.24 µs	125
mut: data set	v3	454 ns	472 ns	702 ns	451 ns	354
mut: data set	gpu	109 ns	112 ns	152 ns	107 ns	1481
mut: position set	v3	1.22 µs	1.25 µs	1.36 µs	1.20 µs	132
mut: position set	gpu	50.7 ns	53.8 ns	102 ns	49.4 ns	3150
mut: select + unselect	v3	2.54 µs	2.71 µs	3.24 µs	2.46 µs	57
mut: select + unselect	gpu	670 ns	733 ns	1.29 µs	619 ns	697576

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

sweep: nodes()

167819×

sweep: edges()

304501×

sweep: elements()

29945×

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

9.4×

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

13×

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

35×

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

59×

sweep: filter(fn)

1.68×

1 ns10 ns100 ns1 µs10 µs100 µs1 ms

All stats (table view)

group	bench	p50	avg	p99	min	samples
sweep: nodes()	v3	413 µs	425 µs	548 µs	407 µs	1651
sweep: nodes()	gpu	2.46 ns	2.68 ns	5.61 ns	2.40 ns	61454
sweep: edges()	v3	781 µs	808 µs	1.19 ms	729 µs	864
sweep: edges()	gpu	2.56 ns	2.66 ns	4.94 ns	2.40 ns	62241
sweep: elements()	v3	144 µs	151 µs	230 µs	142 µs	4602
sweep: elements()	gpu	4.80 ns	4.98 ns	11.4 ns	4.60 ns	33808
sweep: all nodes ($("node") vs filter({group}))	v3	550 µs	574 µs	854 µs	536 µs	1215
sweep: all nodes ($("node") vs filter({group}))	gpu	58.5 µs	73.5 µs	184 µs	54.3 µs	9416
sweep: selected ($(":selected") vs filter({selected}))	v3	345 µs	354 µs	564 µs	333 µs	1971
sweep: selected ($(":selected") vs filter({selected}))	gpu	26.5 µs	26.8 µs	27.8 µs	26.1 µs	12
sweep: selected nodes ($("node:selected") vs filter({group, selected}))	v3	332 µs	342 µs	578 µs	324 µs	2043
sweep: selected nodes ($("node:selected") vs filter({group, selected}))	gpu	9.53 µs	9.63 µs	10.1 µs	9.41 µs	14
sweep: nodes(":selected") vs nodes({selected})	v3	532 µs	550 µs	913 µs	524 µs	1273
sweep: nodes(":selected") vs nodes({selected})	gpu	9.04 µs	9.08 µs	9.41 µs	8.94 µs	15
sweep: filter(fn)	v3	415 µs	424 µs	593 µs	401 µs	1650
sweep: filter(fn)	gpu	246 µs	265 µs	553 µs	240 µs	2621

mutators N=2,000 · 18.0 sv4 (gpu) v3

mut-bulk: select + unselect

7.1×

mut-bulk: hide + show

283×

mut-bulk: lock + unlock

52×

mut-bulk: positions(obj)

105×

mut-bulk: positions(fn)

99×

mut-bulk: shift

100×

mut-bulk: data set

14×

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

6.2×

10 µs100 µs1 ms10 ms100 ms

All stats (table view)

group	bench	p50	avg	p99	min	samples
mut-bulk: select + unselect	v3	1.89 ms	1.94 ms	3.10 ms	1.75 ms	342
mut-bulk: select + unselect	gpu	267 µs	276 µs	397 µs	259 µs	2543
mut-bulk: hide + show	v3	60.3 ms	66.6 ms	70.7 ms	59.3 ms	12
mut-bulk: hide + show	gpu	213 µs	221 µs	391 µs	207 µs	3071
mut-bulk: lock + unlock	v3	2.00 ms	2.08 ms	3.50 ms	1.94 ms	332
mut-bulk: lock + unlock	gpu	38.5 µs	38.9 µs	39.3 µs	38.1 µs	12
mut-bulk: positions(obj)	v3	2.17 ms	2.25 ms	4.65 ms	2.09 ms	303
mut-bulk: positions(obj)	gpu	20.6 µs	22.4 µs	36.7 µs	19.6 µs	30758
mut-bulk: positions(fn)	v3	3.45 ms	3.65 ms	5.87 ms	3.32 ms	188
mut-bulk: positions(fn)	gpu	34.7 µs	37.2 µs	88.5 µs	32.8 µs	18292
mut-bulk: shift	v3	3.41 ms	3.61 ms	6.13 ms	3.30 ms	188
mut-bulk: shift	gpu	34.0 µs	36.2 µs	69.8 µs	32.8 µs	18994
mut-bulk: data set	v3	850 µs	893 µs	1.55 ms	837 µs	766
mut-bulk: data set	gpu	62.3 µs	63.6 µs	102 µs	60.0 µs	10640
mut-bulk: remove + re-add (256 nodes + cascade)	v3	41.4 ms	43.4 ms	49.4 ms	40.5 ms	12
mut-bulk: remove + re-add (256 nodes + cascade)	gpu	6.73 ms	6.84 ms	8.05 ms	6.00 ms	96

scenarios N=2,000 · 12.8 sv4 (gpu) v3

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

8.2×

scn: select-all + fit (listeners on)

26×

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

16×

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

6.7×

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

1.78×

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	307 µs	364 µs	1.15 ms	277 µs	1900
scn: explore (expand 2-hop + select + fit)	gpu	37.6 µs	42.6 µs	182 µs	34.1 µs	16039
scn: select-all + fit (listeners on)	v3	23.9 ms	24.0 ms	26.1 ms	22.4 ms	25
scn: select-all + fit (listeners on)	gpu	910 µs	963 µs	1.49 ms	868 µs	699
scn: drag (select band of 100 + 8-step shift, position listener)	v3	1.64 ms	1.70 ms	2.30 ms	1.57 ms	395
scn: drag (select band of 100 + 8-step shift, position listener)	gpu	102 µs	106 µs	176 µs	100 µs	6324
scn: edit (remove + re-add 256 nodes + cascade, add/remove listeners)	v3	47.2 ms	47.2 ms	49.4 ms	44.1 ms	11
scn: edit (remove + re-add 256 nodes + cascade, add/remove listeners)	gpu	7.09 ms	7.17 ms	8.65 ms	6.33 ms	89
scn: refresh (bulk data write + mapped labels + filter + fit, data listener)	v3	7.53 ms	8.10 ms	12.7 ms	7.00 ms	80
scn: refresh (bulk data write + mapped labels + filter + fit, data listener)	gpu	4.22 ms	4.35 ms	7.10 ms	3.73 ms	148

traversal N=2,000 · 40.1 sv4 (gpu) v3

trav: neighborhood()

4.0×

trav: closedNeighborhood()

2.5×

trav: connectedEdges()

2.6×

trav: outgoers()

5.3×

trav: incomers()

5.1×

trav: successors()

757×

trav: predecessors()

736×

trav: band neighborhood()

4.9×

trav: band connectedEdges()

1.42×

trav: band roots()

22×

trav: band sources()

1.60×

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	4.04 µs	4.04 µs	5.03 µs	3.38 µs	38
trav: neighborhood()	gpu	1.02 µs	1.08 µs	2.15 µs	842 ns	153
trav: closedNeighborhood()	v3	5.55 µs	5.65 µs	6.18 µs	5.18 µs	26
trav: closedNeighborhood()	gpu	2.20 µs	2.37 µs	3.01 µs	2.12 µs	67
trav: connectedEdges()	v3	1.04 µs	1.15 µs	2.18 µs	986 ns	142
trav: connectedEdges()	gpu	397 ns	415 ns	705 ns	385 ns	405
trav: outgoers()	v3	2.32 µs	2.47 µs	3.05 µs	2.23 µs	65
trav: outgoers()	gpu	439 ns	523 ns	2.01 µs	429 ns	317
trav: incomers()	v3	2.26 µs	2.41 µs	3.07 µs	2.19 µs	66
trav: incomers()	gpu	442 ns	530 ns	2.04 µs	431 ns	311
trav: successors()	v3	428 ms	429 ms	433 ms	424 ms	12
trav: successors()	gpu	565 µs	593 µs	1.90 ms	545 µs	1151
trav: predecessors()	v3	432 ms	432 ms	439 ms	425 ms	12
trav: predecessors()	gpu	587 µs	623 µs	1.93 ms	565 µs	1093
trav: band neighborhood()	v3	182 µs	197 µs	356 µs	178 µs	3531
trav: band neighborhood()	gpu	37.3 µs	38.2 µs	39.9 µs	37.0 µs	12
trav: band connectedEdges()	v3	24.0 µs	24.1 µs	24.5 µs	23.4 µs	12
trav: band connectedEdges()	gpu	16.9 µs	17.6 µs	19.2 µs	16.5 µs	12
trav: band roots()	v3	109 µs	117 µs	162 µs	102 µs	5905
trav: band roots()	gpu	4.90 µs	4.94 µs	5.17 µs	4.79 µs	31
trav: band sources()	v3	7.01 µs	7.03 µs	7.36 µs	6.55 µs	21
trav: band sources()	gpu	4.39 µs	4.42 µs	4.73 µs	4.30 µs	34
trav: band connectedNodes()	v3	14.1 µs	14.4 µs	14.3 µs	13.8 µs	12
trav: band connectedNodes()	gpu	4.96 µs	5.02 µs	5.31 µs	4.85 µs	30

algorithms N=2,000 · 1.6 minv4 (gpu) v3

algo: bfs (whole graph)

62×

algo: dfs (whole graph)

61×

algo: dijkstra + pathTo

57×

algo: aStar

3.4×

algo: bellmanFord

20×

algo: kruskal

27×

algo: tarjan SCC

38×

algo: hopcroft-tarjan biconnected

26×

algo: hierholzer (directed)

3.0×

algo: betweennessCentrality (unweighted)

11×

algo: degreeCentralityNormalized

38×

algo: closenessCentrality (one root)

74×

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	56.7 ms	55.7 ms	60.2 ms	50.7 ms	9
algo: bfs (whole graph)	gpu	918 µs	940 µs	1.53 ms	854 µs	737
algo: dfs (whole graph)	v3	50.2 ms	51.0 ms	51.9 ms	49.6 ms	10
algo: dfs (whole graph)	gpu	828 µs	856 µs	1.26 ms	803 µs	796
algo: dijkstra + pathTo	v3	52.6 ms	54.1 ms	58.4 ms	52.1 ms	9
algo: dijkstra + pathTo	gpu	931 µs	947 µs	1.18 ms	910 µs	725
algo: aStar	v3	1.74 ms	1.80 ms	2.56 ms	1.65 ms	384
algo: aStar	gpu	507 µs	521 µs	695 µs	496 µs	1339
algo: bellmanFord	v3	233 ms	230 ms	251 ms	211 ms	12
algo: bellmanFord	gpu	11.7 ms	12.0 ms	14.7 ms	11.6 ms	54
algo: kruskal	v3	13.8 ms	14.7 ms	18.4 ms	12.9 ms	43
algo: kruskal	gpu	503 µs	527 µs	870 µs	468 µs	1304
algo: tarjan SCC	v3	24.4 ms	24.1 ms	27.6 ms	21.4 ms	25
algo: tarjan SCC	gpu	647 µs	816 µs	5.34 ms	615 µs	827
algo: hopcroft-tarjan biconnected	v3	87.8 ms	88.9 ms	92.9 ms	83.1 ms	12
algo: hopcroft-tarjan biconnected	gpu	3.43 ms	3.54 ms	4.55 ms	3.34 ms	191
algo: hierholzer (directed)	v3	23.2 ms	24.0 ms	26.9 ms	22.0 ms	25
algo: hierholzer (directed)	gpu	7.75 ms	7.93 ms	9.50 ms	7.64 ms	78
algo: betweennessCentrality (unweighted)	v3	4.07 s	4.08 s	4.15 s	4.02 s	12
algo: betweennessCentrality (unweighted)	gpu	359 ms	380 ms	449 ms	351 ms	12
algo: degreeCentralityNormalized	v3	15.1 ms	15.7 ms	18.5 ms	13.9 ms	40
algo: degreeCentralityNormalized	gpu	401 µs	411 µs	656 µs	386 µs	1694
algo: closenessCentrality (one root)	v3	60.3 ms	62.9 ms	68.1 ms	58.5 ms	12
algo: closenessCentrality (one root)	gpu	816 µs	836 µs	1.21 ms	782 µs	837
algo: degreeCentrality (one root)	v3	7.39 µs	7.18 µs	7.69 µs	6.59 µs	20
algo: degreeCentrality (one root)	gpu	244 ns	249 ns	301 ns	239 ns	678

algorithms N=500 · 6.9 minv4 (gpu) v3

algo: bfs (whole graph)

65×

algo: dfs (whole graph)

63×

algo: dijkstra + pathTo

55×

algo: aStar

3.6×

algo: bellmanFord

21×

algo: kruskal

25×

algo: tarjan SCC

38×

algo: hopcroft-tarjan biconnected

26×

algo: hierholzer (directed)

5.4×

algo: betweennessCentrality (unweighted)

11×

algo: degreeCentralityNormalized

40×

algo: closenessCentrality (one root)

75×

algo: degreeCentrality (one root)

32×

algo: pageRank (20 iters)

1.92×

algo: closenessCentrality (weighted, one root)

57×

algo: floydWarshall

1.60×

algo: closenessCentralityNormalized

1.68×

algo: markovClustering

1.10×

algo: hierarchicalClustering (threshold)

2.5×

algo: kMeans (fixed centroids)

3.9×

algo: kMedoids

8.7×

algo: fuzzyCMeans

8.6×

algo: affinityPropagation

5.2×

algo: kargerStein

1.03×

algo: betweennessCentrality (weighted)

9.0×

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	12.8 ms	13.6 ms	19.5 ms	11.4 ms	47
algo: bfs (whole graph)	gpu	198 µs	212 µs	397 µs	193 µs	3305
algo: dfs (whole graph)	v3	11.8 ms	12.8 ms	16.7 ms	11.4 ms	51
algo: dfs (whole graph)	gpu	186 µs	191 µs	248 µs	183 µs	3663
algo: dijkstra + pathTo	v3	12.4 ms	13.4 ms	16.8 ms	12.0 ms	48
algo: dijkstra + pathTo	gpu	227 µs	236 µs	336 µs	221 µs	2966
algo: aStar	v3	434 µs	443 µs	522 µs	424 µs	1580
algo: aStar	gpu	119 µs	122 µs	165 µs	116 µs	5721
algo: bellmanFord	v3	17.3 ms	17.4 ms	18.1 ms	17.0 ms	36
algo: bellmanFord	gpu	805 µs	836 µs	1.91 ms	800 µs	832
algo: kruskal	v3	3.31 ms	3.58 ms	6.88 ms	3.19 ms	191
algo: kruskal	gpu	132 µs	140 µs	242 µs	127 µs	4954
algo: tarjan SCC	v3	5.63 ms	6.04 ms	9.10 ms	5.45 ms	111
algo: tarjan SCC	gpu	147 µs	156 µs	264 µs	144 µs	4490
algo: hopcroft-tarjan biconnected	v3	21.6 ms	22.2 ms	25.4 ms	20.0 ms	28
algo: hopcroft-tarjan biconnected	gpu	818 µs	838 µs	1.20 ms	792 µs	831
algo: hierholzer (directed)	v3	4.33 ms	4.60 ms	7.79 ms	4.20 ms	145
algo: hierholzer (directed)	gpu	801 µs	825 µs	1.18 ms	788 µs	835
algo: betweennessCentrality (unweighted)	v3	264 ms	265 ms	274 ms	259 ms	12
algo: betweennessCentrality (unweighted)	gpu	24.2 ms	24.3 ms	25.7 ms	23.0 ms	25
algo: degreeCentralityNormalized	v3	3.69 ms	4.00 ms	7.49 ms	3.57 ms	171
algo: degreeCentralityNormalized	gpu	92.7 µs	95.7 µs	148 µs	89.5 µs	7294
algo: closenessCentrality (one root)	v3	14.7 ms	15.8 ms	20.1 ms	14.0 ms	40
algo: closenessCentrality (one root)	gpu	197 µs	206 µs	296 µs	191 µs	3396
algo: degreeCentrality (one root)	v3	7.72 µs	7.56 µs	8.13 µs	6.91 µs	19
algo: degreeCentrality (one root)	gpu	241 ns	247 ns	375 ns	230 ns	688
algo: pageRank (20 iters)	v3	2.76 ms	2.90 ms	6.35 ms	2.69 ms	237
algo: pageRank (20 iters)	gpu	1.44 ms	1.67 ms	3.78 ms	1.42 ms	414
algo: closenessCentrality (weighted, one root)	v3	14.3 ms	15.5 ms	18.8 ms	13.8 ms	40
algo: closenessCentrality (weighted, one root)	gpu	249 µs	256 µs	334 µs	241 µs	2734
algo: floydWarshall	v3	247 ms	273 ms	379 ms	244 ms	12
algo: floydWarshall	gpu	154 ms	159 ms	209 ms	139 ms	12
algo: closenessCentralityNormalized	v3	274 ms	281 ms	297 ms	270 ms	12
algo: closenessCentralityNormalized	gpu	163 ms	165 ms	166 ms	162 ms	12
algo: markovClustering	v3	4.14 s	4.38 s	4.66 s	4.08 s	12
algo: markovClustering	gpu	3.76 s	3.76 s	3.77 s	3.74 s	12
algo: hierarchicalClustering (threshold)	v3	35.3 ms	35.4 ms	36.0 ms	34.6 ms	16
algo: hierarchicalClustering (threshold)	gpu	14.2 ms	14.8 ms	20.1 ms	13.8 ms	42
algo: kMeans (fixed centroids)	v3	9.12 ms	9.20 ms	9.86 ms	8.90 ms	72
algo: kMeans (fixed centroids)	gpu	2.32 ms	2.39 ms	3.42 ms	2.28 ms	289
algo: kMedoids	v3	363 ms	290 ms	376 ms	70.4 ms	12
algo: kMedoids	gpu	41.9 ms	42.6 ms	57.5 ms	22.2 ms	13
algo: fuzzyCMeans	v3	78.1 ms	79.6 ms	85.0 ms	77.1 ms	12
algo: fuzzyCMeans	gpu	9.11 ms	9.18 ms	9.69 ms	8.80 ms	72
algo: affinityPropagation	v3	209 ms	211 ms	216 ms	202 ms	12
algo: affinityPropagation	gpu	40.0 ms	41.4 ms	45.1 ms	39.0 ms	10
algo: kargerStein	v3	317 ms	322 ms	334 ms	313 ms	12
algo: kargerStein	gpu	308 ms	309 ms	317 ms	300 ms	12
algo: betweennessCentrality (weighted)	v3	720 ms	725 ms	738 ms	712 ms	12
algo: betweennessCentrality (weighted)	gpu	79.8 ms	81.9 ms	83.6 ms	78.3 ms	12

mappers N=2,000 · 8.2 s

mapper: bulk write, color — cpu eval

2.77 ms

mapper: bulk write, color — gpu path

175 µs

mapper: bulk write, color + size — cpu eval

3.14 ms

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

2.79 ms

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

3.95 ms

100 µs1 ms10 ms

All stats (table view)

group	bench	p50	avg	p99	min	samples
mapper: bulk write, color	cpu eval	2.77 ms	2.86 ms	4.59 ms	2.66 ms	239
mapper: bulk write, color	gpu path	175 µs	182 µs	270 µs	169 µs	3856
mapper: bulk write, color + size	cpu eval	3.14 ms	3.17 ms	3.63 ms	3.02 ms	215
mapper: bulk write, color + size	gpu color / cpu size	2.79 ms	2.85 ms	3.70 ms	2.68 ms	242
mapper: bulk write, mapped labels (round-5 path)	label refresh	3.95 ms	4.16 ms	6.80 ms	3.48 ms	163

layouts N=2,000 · 30.9 sv4 (gpu) v3

154×

18×

57×

12×

26×

37×

11×

113 ms

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

111 ms

55.9 ms

360 µs

31.1 µs

529 ns

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

39.9 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	11.6 ms	12.3 ms	18.7 ms	9.93 ms	51
layout: grid	gpu	75.5 µs	81.3 µs	359 µs	72.9 µs	7877
layout: preset	v3	5.83 ms	6.44 ms	11.8 ms	5.51 ms	103
layout: preset	gpu	329 µs	349 µs	773 µs	292 µs	1874
layout: preset (fn form)	v3	5.56 ms	5.90 ms	9.25 ms	5.22 ms	113
layout: preset (fn form)	gpu	96.9 µs	108 µs	391 µs	94.1 µs	6387
layout: circle	v3	9.58 ms	10.1 ms	14.8 ms	9.01 ms	65
layout: circle	gpu	800 µs	896 µs	2.93 ms	762 µs	755
layout: concentric	v3	31.0 ms	31.9 ms	37.9 ms	28.2 ms	18
layout: concentric	gpu	1.21 ms	1.40 ms	4.91 ms	1.19 ms	484
layout: breadthfirst	v3	410 ms	413 ms	437 ms	393 ms	12
layout: breadthfirst	gpu	11.2 ms	12.2 ms	18.8 ms	10.4 ms	51
layout: random	v3	5.33 ms	5.64 ms	9.49 ms	5.17 ms	118
layout: random	gpu	497 µs	549 µs	1.49 ms	478 µs	1254
layout: force (CPU executor, 20 iterations)	gpu	113 ms	129 ms	143 ms	103 ms	12
layout: force seed — spectral vs scatter (20 iterations)	spectral (landmark MDS, the default)	111 ms	111 ms	119 ms	103 ms	12
layout: force seed — spectral vs scatter (20 iterations)	scatter	55.9 ms	56.5 ms	56.7 ms	55.5 ms	9
layout: plumbing (layoutPositions vs setPositions)	finisher	360 µs	394 µs	861 µs	350 µs	1756
layout: plumbing (layoutPositions vs setPositions)	bulk	31.1 µs	37.8 µs	66.8 µs	29.3 µs	16741
layout: contract fixed cost (empty impl vs bulk placement)	wrapper only	529 ns	978 ns	4.00 µs	511 ns	166
layout: contract fixed cost (empty impl vs bulk placement)	wrapper + bulk write	39.9 µs	41.9 µs	48.8 µs	37.4 µs	12
layout: scope (eles.layout, 10% of nodes)	v3	969 µs	1.05 ms	2.16 ms	948 µs	648
layout: scope (eles.layout, 10% of nodes)	gpu	102 µs	118 µs	262 µs	99.0 µs	5796

style N=2,000 · 34.7 sv4 (gpu) v3

style: sheet swap (compile + applyAll)

1.14×

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

32.6 µs

style: compile vs compile + apply — compile + apply

10.5 ms

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

1.61×

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

10.7 ms

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

11.3 ms

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

10.5 ms

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

10.6 ms

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

10.8 ms

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

14.3 ms

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

6.77 µs

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

34.4 µs

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

573 µs

style: read one prop (color)

1.48×

style: read one prop (width)

2.1×

style: numericStyle

2.00×

style: renderedStyle (one prop)

14×

style: whole-object style()

6.2×

style: bypass set (ele.style name,value)

3.2×

style: bypass set + removeStyle round-trip

4.3×

style: select+unselect 256-band under the default sheet — bypass punch-out — no bypasses (diff path; the O(0) gate row)

35.2 µs

style: select+unselect 256-band under the default sheet — bypass punch-out — 128 of 256 bypassed (punched out to merged writes)

343 µs

style: sheet swap (applyAll) — 0 vs 128 bypasses — no bypasses

11.0 ms

style: sheet swap (applyAll) — 0 vs 128 bypasses — 128 bypassed

11.5 ms

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	12.4 ms	12.8 ms	16.7 ms	11.7 ms	47
style: sheet swap (compile + applyAll)	gpu	10.8 ms	11.2 ms	13.6 ms	10.5 ms	58
style: compile vs compile + apply	compile only (in batch)	32.6 µs	35.5 µs	41.5 µs	28.8 µs	12
style: compile vs compile + apply	compile + apply	10.5 ms	10.8 ms	13.5 ms	10.3 ms	60
style: first apply on add (256-node band)	v3	1.32 ms	1.56 ms	6.79 ms	1.10 ms	432
style: first apply on add (256-node band)	gpu	820 µs	928 µs	2.84 ms	785 µs	744
style: applyAll, flat vs compound (100 parents — the 14.6 partition)	flat	10.7 ms	10.9 ms	13.0 ms	10.4 ms	61
style: applyAll, flat vs compound (100 parents — the 14.6 partition)	compound	11.3 ms	11.4 ms	13.8 ms	10.9 ms	57
style: applyAll — constant vs state-case (the 57.1d partition)	constant sheet	10.5 ms	10.7 ms	11.5 ms	10.3 ms	60
style: applyAll — constant vs state-case (the 57.1d partition)	state-case sheet	10.6 ms	10.8 ms	13.1 ms	10.3 ms	59
style: applyAll — state-case vs data-case (what the partition avoids)	state-case sheet	10.8 ms	11.0 ms	13.0 ms	10.5 ms	60
style: applyAll — state-case vs data-case (what the partition avoids)	data-case sheet	14.3 ms	14.5 ms	15.7 ms	13.9 ms	42
style: select+unselect 256-band — constant vs state-case sheet	constant sheet (skips restyle)	6.77 µs	6.81 µs	6.99 µs	6.70 µs	20
style: select+unselect 256-band — constant vs state-case sheet	paint-only state sheet (round-61 diff path)	34.4 µs	35.6 µs	62.5 µs	32.7 µs	18358
style: select+unselect 256-band — constant vs state-case sheet	state-case sheet (all channels; full write)	573 µs	605 µs	1.07 ms	554 µs	1146
style: read one prop (color)	v3	91.0 ns	95.7 ns	172 ns	81.0 ns	2078250
style: read one prop (color)	gpu	61.3 ns	67.8 ns	122 ns	59.1 ns	2505
style: read one prop (width)	v3	109 ns	120 ns	521 ns	98.0 ns	1938820
style: read one prop (width)	gpu	51.6 ns	56.4 ns	307 ns	49.4 ns	3021
style: numericStyle	v3	98.0 ns	105 ns	367 ns	83.0 ns	2018194
style: numericStyle	gpu	49.0 ns	52.0 ns	116 ns	47.8 ns	3283
style: renderedStyle (one prop)	v3	903 ns	971 ns	1.96 µs	851 ns	540257
style: renderedStyle (one prop)	gpu	64.5 ns	68.1 ns	214 ns	61.4 ns	2511
style: whole-object style()	v3	87.3 µs	92.4 µs	182 µs	83.9 µs	7532
style: whole-object style()	gpu	14.0 µs	15.4 µs	52.2 µs	13.4 µs	43844
style: bypass set (ele.style name,value)	v3	9.97 µs	10.0 µs	10.1 µs	9.68 µs	11
style: bypass set (ele.style name,value)	gpu	3.15 µs	3.44 µs	4.34 µs	3.03 µs	43
style: bypass set + removeStyle round-trip	v3	21.7 µs	24.1 µs	48.6 µs	20.8 µs	27716
style: bypass set + removeStyle round-trip	gpu	5.00 µs	5.25 µs	6.37 µs	4.75 µs	28
style: select+unselect 256-band under the default sheet — bypass punch-out	no bypasses (diff path; the O(0) gate row)	35.2 µs	36.6 µs	68.7 µs	34.1 µs	18163
style: select+unselect 256-band under the default sheet — bypass punch-out	128 of 256 bypassed (punched out to merged writes)	343 µs	359 µs	563 µs	327 µs	1940
style: sheet swap (applyAll) — 0 vs 128 bypasses	no bypasses	11.0 ms	11.1 ms	12.8 ms	10.7 ms	56
style: sheet swap (applyAll) — 0 vs 128 bypasses	128 bypassed	11.5 ms	11.7 ms	14.5 ms	11.0 ms	53

style-bundle N=2,000 · 19.6 sv4 (gpu) v3

style( 'background-color' ) (was #4 in the switch)

1.14×

style( 'border-width' ) (was #6 in the switch)

1.19×

style( 'text-wrap' ) (was #73 in the switch)

1.67×

style( 'text-max-width' ) (was #74 in the switch)

1.69×

style( 'taxi-radius' ) (was #142 in the switch)

1.22×

style( 'target-distance-from-node' ) (was #150 in the switch)

1.22×

whole-object style() (gpu-only: v3 returns a different shape) — gpu node

13.4 µs

whole-object style() (gpu-only: v3 returns a different shape) — gpu edge

14.5 µs

colour-valued vs numeric reads (the string build) — gpu background-color (colour)

70.2 ns

colour-valued vs numeric reads (the string build) — gpu border-color (colour)

74.0 ns

colour-valued vs numeric reads (the string build) — gpu border-width (number)

63.2 ns

colour-valued vs numeric reads (the string build) — gpu width (number)

59.6 ns

the getters that ride readProp — gpu numericStyle( width )

60.0 ns

the getters that ride readProp — gpu effectiveOpacity()

20.7 ns

the getters that ride readProp — gpu renderedStyle( width )

78.6 ns

10 ns100 ns1 µs10 µs100 µs

All stats (table view)

group	bench	p50	avg	p99	min	samples
style( 'background-color' ) (was #4 in the switch)	v3	59.6 ns	61.0 ns	82.5 ns	56.5 ns	2808
style( 'background-color' ) (was #4 in the switch)	gpu	52.3 ns	54.9 ns	95.5 ns	49.8 ns	3105
style( 'border-width' ) (was #6 in the switch)	v3	60.3 ns	62.4 ns	97.5 ns	58.5 ns	2745
style( 'border-width' ) (was #6 in the switch)	gpu	50.6 ns	51.8 ns	64.2 ns	49.6 ns	3292
style( 'text-wrap' ) (was #73 in the switch)	v3	61.3 ns	64.1 ns	112 ns	60.1 ns	2661
style( 'text-wrap' ) (was #73 in the switch)	gpu	36.6 ns	38.1 ns	65.3 ns	35.6 ns	4499
style( 'text-max-width' ) (was #74 in the switch)	v3	58.6 ns	59.9 ns	77.8 ns	57.1 ns	2853
style( 'text-max-width' ) (was #74 in the switch)	gpu	34.6 ns	35.7 ns	58.0 ns	33.3 ns	4793
style( 'taxi-radius' ) (was #142 in the switch)	v3	59.6 ns	61.3 ns	79.6 ns	57.8 ns	2801
style( 'taxi-radius' ) (was #142 in the switch)	gpu	48.9 ns	52.3 ns	75.6 ns	47.4 ns	3253
style( 'target-distance-from-node' ) (was #150 in the switch)	v3	59.5 ns	61.0 ns	75.5 ns	57.2 ns	2809
style( 'target-distance-from-node' ) (was #150 in the switch)	gpu	48.7 ns	50.2 ns	64.8 ns	47.5 ns	3407
whole-object style() (gpu-only: v3 returns a different shape)	gpu node	13.4 µs	14.4 µs	33.9 µs	12.7 µs	46925
whole-object style() (gpu-only: v3 returns a different shape)	gpu edge	14.5 µs	15.8 µs	40.6 µs	13.8 µs	42830
colour-valued vs numeric reads (the string build)	gpu background-color (colour)	70.2 ns	72.0 ns	88.9 ns	69.1 ns	2367
colour-valued vs numeric reads (the string build)	gpu border-color (colour)	74.0 ns	76.5 ns	115 ns	73.2 ns	2234
colour-valued vs numeric reads (the string build)	gpu border-width (number)	63.2 ns	65.2 ns	82.8 ns	61.9 ns	2618
colour-valued vs numeric reads (the string build)	gpu width (number)	59.6 ns	60.9 ns	77.9 ns	57.7 ns	2802
the getters that ride readProp	gpu numericStyle( width )	60.0 ns	61.4 ns	75.2 ns	58.9 ns	2784
the getters that ride readProp	gpu effectiveOpacity()	20.7 ns	21.2 ns	32.9 ns	19.8 ns	8071
the getters that ride readProp	gpu renderedStyle( width )	78.6 ns	79.5 ns	90.0 ns	77.0 ns	2149

load N=2,000 · 19.5 sv4 (gpu) v3

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

6.2×

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

21.3 ms

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

12.5 ms

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

12.1 ms

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

304 µs

load: conversion — toColumnarElements

708 µs

load: conversion — serializeElements (from columnar)

204 µs

load: conversion — serializeElements (from defs)

886 µs

load: conversion — deserializeElements

4.63 µs

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

492 µs

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

737 µs

load: cy.json()

2.7×

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

1.33×

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

2.61 ms

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

2.28 ms

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

2.26 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	132 ms	143 ms	174 ms	127 ms	12
load: init from the definition form (construct + dispose)	gpu	21.3 ms	23.3 ms	29.1 ms	19.8 ms	25
load: v4 ingest forms (construct + dispose)	definition form	21.3 ms	21.9 ms	25.0 ms	18.9 ms	28
load: v4 ingest forms (construct + dispose)	columnar form	12.5 ms	13.6 ms	24.9 ms	12.1 ms	46
load: v4 ingest forms (construct + dispose)	wire buffer	12.1 ms	13.4 ms	24.4 ms	11.7 ms	47
load: v4 ingest forms (construct + dispose)	def clone alone (control)	304 µs	640 µs	13.6 ms	211 µs	1037
load: conversion	toColumnarElements	708 µs	780 µs	1.43 ms	700 µs	885
load: conversion	serializeElements (from columnar)	204 µs	379 µs	1.04 ms	171 µs	1824
load: conversion	serializeElements (from defs)	886 µs	947 µs	1.87 ms	861 µs	733
load: conversion	deserializeElements	4.63 µs	4.38 µs	6.32 µs	3.32 µs	30
load: export from a live graph	cy.serialize() (wire)	492 µs	563 µs	2.00 ms	484 µs	1221
load: export from a live graph	cy.json() (v4)	737 µs	1.08 ms	6.54 ms	704 µs	615
load: cy.json()	v3	2.00 ms	2.32 ms	5.24 ms	1.94 ms	293
load: cy.json()	gpu	742 µs	823 µs	2.70 ms	733 µs	845
load: add a 256-node band (add + remove)	v3	3.58 ms	4.25 ms	10.6 ms	3.22 ms	155
load: add a 256-node band (add + remove)	gpu	2.69 ms	2.96 ms	7.24 ms	2.13 ms	227
load: add a 256-node band, v4 forms	definition form	2.61 ms	2.64 ms	4.53 ms	2.06 ms	259
load: add a 256-node band, v4 forms	columnar form	2.28 ms	2.34 ms	2.99 ms	1.88 ms	288
load: add a 256-node band, v4 forms	wire buffer	2.26 ms	2.33 ms	3.03 ms	1.88 ms	294

spatial N=2,000 · 37.6 sv4 (gpu) v3

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

20.5 µs

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

21.1 µs

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

45.9 µs

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

44.3 µs

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

87.5 µs

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

233 µs

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

829 µs

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

689 µs

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

99.6 µs

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

228 µs

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

827 µs

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

811 µs

pick: background miss (full highWater walk) — gpu

45.4 µs

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

45.2 µs

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

125 µs

box: elementsInBox by covered fraction — 10%

148 µs

box: elementsInBox by covered fraction — 50%

197 µs

box: elementsInBox by covered fraction — 100%

344 µs

box: label containment (boxSelectionIncludesLabels) — off

231 µs

box: label containment (boxSelectionIncludesLabels) — on

261 µs

box: overlap vs contain (straight edges) — contain

241 µs

box: overlap vs contain (straight edges) — overlap

473 µs

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

1.24 ms

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

2.13 ms

bounds: whole-graph boundingBox()

6.0×

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

1.26 ms

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

693 µs

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

31×

bounds: getFitViewport() (compute without committing)

32×

bounds: boundingBoxAt (v4 only) — gpu

2.71 ms

bounds: one node boundingBox()

1.62×

bounds: labelBoundingBox (v4 only) — gpu

419 ns

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

74.4 µs

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

148 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	20.5 µs	21.2 µs	40.7 µs	20.0 µs	32390
pick: hit (the walk; one inside-test)	zoom 2	21.1 µs	21.6 µs	21.9 µs	20.8 µs	12
pick: miss walk — exact vs 24 px touch halo (57.9)	exact	45.9 µs	46.7 µs	49.7 µs	45.5 µs	12
pick: miss walk — exact vs 24 px touch halo (57.9)	padPx 24	44.3 µs	46.1 µs	87.5 µs	42.3 µs	15082
pick: shape inside-test, 2000 coincident misses	ellipse	87.5 µs	91.4 µs	190 µs	85.1 µs	7640
pick: shape inside-test, 2000 coincident misses	star	233 µs	244 µs	436 µs	225 µs	2858
pick: shape inside-test, 2000 coincident misses	round-hexagon	829 µs	851 µs	1.28 ms	823 µs	821
pick: shape inside-test, 2000 coincident misses	barrel	689 µs	712 µs	1.10 ms	680 µs	978
pick: shape inside-test, 2000 coincident misses	cut-rectangle	99.6 µs	103 µs	181 µs	95.6 µs	6809
pick: shape inside-test, 2000 coincident misses	polygon	228 µs	238 µs	407 µs	212 µs	2935
pick: round-hexagon inside-test by zoom (not affine-invariant)	zoom 1	827 µs	830 µs	1.08 ms	787 µs	842
pick: round-hexagon inside-test by zoom (not affine-invariant)	zoom 2	811 µs	821 µs	989 µs	785 µs	852
pick: background miss (full highWater walk)	gpu	45.4 µs	45.8 µs	46.0 µs	45.0 µs	12
pick: label box in the scan (text-events)	text-events: no	45.2 µs	45.5 µs	45.9 µs	44.9 µs	12
pick: label box in the scan (text-events)	text-events: yes	125 µs	129 µs	210 µs	121 µs	5361
box: elementsInBox by covered fraction	10%	148 µs	153 µs	248 µs	147 µs	4562
box: elementsInBox by covered fraction	50%	197 µs	201 µs	249 µs	191 µs	3442
box: elementsInBox by covered fraction	100%	344 µs	356 µs	814 µs	325 µs	1954
box: label containment (boxSelectionIncludesLabels)	off	231 µs	238 µs	353 µs	227 µs	2929
box: label containment (boxSelectionIncludesLabels)	on	261 µs	270 µs	354 µs	257 µs	2538
box: overlap vs contain (straight edges)	contain	241 µs	251 µs	603 µs	235 µs	2778
box: overlap vs contain (straight edges)	overlap	473 µs	497 µs	882 µs	409 µs	1404
box: overlap vs contain (curved edges — the segment walk)	contain	1.24 ms	1.26 ms	1.73 ms	1.21 ms	554
box: overlap vs contain (curved edges — the segment walk)	overlap	2.13 ms	2.17 ms	2.89 ms	2.04 ms	320
bounds: whole-graph boundingBox()	v3	7.54 ms	7.73 ms	9.14 ms	7.16 ms	85
bounds: whole-graph boundingBox()	gpu	1.26 ms	1.29 ms	1.70 ms	1.21 ms	537
bounds: boundingBox, labels in vs out (v4)	includeLabels: true (default)	1.26 ms	1.29 ms	1.87 ms	1.21 ms	539
bounds: boundingBox, labels in vs out (v4)	includeLabels: false	693 µs	713 µs	1.08 ms	667 µs	981
bounds: cy.fit() (whole-graph scan)	v3	7.46 ms	7.57 ms	8.46 ms	7.23 ms	89
bounds: cy.fit() (whole-graph scan)	gpu	237 µs	245 µs	348 µs	233 µs	2841
bounds: getFitViewport() (compute without committing)	v3	7.40 ms	7.60 ms	9.16 ms	7.20 ms	88
bounds: getFitViewport() (compute without committing)	gpu	233 µs	241 µs	328 µs	226 µs	2894
bounds: boundingBoxAt (v4 only)	gpu	2.71 ms	2.78 ms	3.62 ms	2.65 ms	244
bounds: one node boundingBox()	v3	1.61 µs	1.63 µs	2.13 µs	1.58 µs	100
bounds: one node boundingBox()	gpu	996 ns	1.01 µs	1.19 µs	981 ns	161
bounds: labelBoundingBox (v4 only)	gpu	419 ns	450 ns	827 ns	409 ns	371
bounds: compound + taxi conservative scan (round 54, v4 only)	warm (curve bbs memoized)	74.4 µs	78.0 µs	186 µs	73.8 µs	8930
bounds: compound + taxi conservative scan (round 54, v4 only)	cold (geometry write per call)	148 µs	154 µs	299 µs	141 µs	4537

data N=2,000 · 28.4 sv4 (gpu) v3

data: bulk write, numeric column

19×

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

16×

data: bulk write, one new dictionary entry per pass

16×

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

19×

data: removeData (whole collection)

7.2×

data: read one numeric key

1.35×

data: read one string key (dictionary decode)

1.31×

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 })

17×

query: predicate function (both sides)

8.5×

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

34×

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	1.64 ms	1.68 ms	2.80 ms	1.56 ms	399
data: bulk write, numeric column	gpu	84.3 µs	87.4 µs	129 µs	80.9 µs	8001
data: bulk write, dictionary string column (4 values)	v3	1.64 ms	1.68 ms	2.25 ms	1.60 ms	397
data: bulk write, dictionary string column (4 values)	gpu	101 µs	107 µs	312 µs	94.8 µs	6520
data: bulk write, one new dictionary entry per pass	v3	1.65 ms	1.73 ms	2.89 ms	1.58 ms	395
data: bulk write, one new dictionary entry per pass	gpu	100 µs	105 µs	179 µs	96.4 µs	6683
data: bulk write, object column (plain-array fallback)	v3	1.66 ms	1.74 ms	2.98 ms	1.61 ms	397
data: bulk write, object column (plain-array fallback)	gpu	84.9 µs	86.6 µs	117 µs	81.0 µs	8079
data: removeData (whole collection)	v3	2.42 ms	2.49 ms	4.21 ms	2.36 ms	275
data: removeData (whole collection)	gpu	336 µs	347 µs	592 µs	326 µs	1971
data: read one numeric key	v3	38.5 ns	40.1 ns	66.1 ns	37.3 ns	4266
data: read one numeric key	gpu	28.6 ns	30.0 ns	43.1 ns	27.8 ns	5695
data: read one string key (dictionary decode)	v3	41.6 ns	43.3 ns	82.9 ns	40.3 ns	3969
data: read one string key (dictionary decode)	gpu	31.7 ns	32.9 ns	70.1 ns	30.6 ns	5223
data: read the whole object	v3	48.0 ns	49.2 ns	72.4 ns	45.7 ns	3493
data: read the whole object	gpu	21.0 ns	22.3 ns	43.5 ns	20.2 ns	7639
query: data equality ([bucket = 3] vs { data: { bucket: 3 } })	v3	656 µs	682 µs	1.14 ms	637 µs	1023
query: data equality ([bucket = 3] vs { data: { bucket: 3 } })	gpu	52.0 µs	56.1 µs	129 µs	46.0 µs	12236
query: data comparison ([weight > 3] vs { gt })	v3	758 µs	779 µs	1.06 ms	751 µs	893
query: data comparison ([weight > 3] vs { gt })	gpu	61.7 µs	71.0 µs	152 µs	57.4 µs	9517
query: two keys AND-ed	v3	791 µs	816 µs	1.31 ms	770 µs	855
query: two keys AND-ed	gpu	68.5 µs	70.9 µs	112 µs	65.7 µs	9538
query: membership ([bucket in ...] vs { in })	v3	1.11 ms	1.15 ms	1.63 ms	1.01 ms	604
query: membership ([bucket in ...] vs { in })	gpu	66.6 µs	70.8 µs	144 µs	62.9 µs	9564
query: predicate function (both sides)	v3	653 µs	667 µs	914 µs	637 µs	1047
query: predicate function (both sides)	gpu	76.6 µs	79.9 µs	150 µs	75.3 µs	8714
query: structural, parents (:parent vs { parent: true })	v3	532 µs	541 µs	719 µs	519 µs	1291
query: structural, parents (:parent vs { parent: true })	gpu	15.5 µs	17.6 µs	23.4 µs	15.4 µs	12
query: structural, childless (:childless vs { parent: false })	v3	802 µs	823 µs	1.28 ms	781 µs	847
query: structural, childless (:childless vs { parent: false })	gpu	60.1 µs	68.0 µs	185 µs	56.5 µs	9861
query: structural, orphans (:orphan vs { child: false })	v3	509 µs	518 µs	730 µs	496 µs	1348
query: structural, orphans (:orphan vs { child: false })	gpu	15.5 µs	16.7 µs	20.4 µs	15.3 µs	12

events N=2,000 · 27.0 sv4 (gpu) v3

emit: position write, no listeners

24×

emit: 1 core listener

8.1×

emit: 1 element (ref-qualified) listener

7.3×

emit: delegated (v3 selector vs v4 predicate)

7.7×

emit: 10 core listeners

6.3×

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

2.2×

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

297 ns

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

834 ns

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

80.2 ns

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

487 ns

anim: start + stop, one element

1.46×

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

1.35 µs

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

839 ns

anim: start + stop, 512-node collection

3.0×

anim: delay() (no channels touched)

1.56×

anim: controls (v4 only) — pause + resume

449 ns

anim: controls (v4 only) — reverse

849 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	1.28 µs	1.30 µs	1.48 µs	1.24 µs	116
emit: position write, no listeners	gpu	52.3 ns	55.2 ns	122 ns	49.6 ns	3088
emit: 1 core listener	v3	1.38 µs	1.40 µs	1.58 µs	1.35 µs	112
emit: 1 core listener	gpu	171 ns	181 ns	274 ns	167 ns	890
emit: 1 element (ref-qualified) listener	v3	1.38 µs	1.44 µs	2.19 µs	1.34 µs	108
emit: 1 element (ref-qualified) listener	gpu	189 ns	205 ns	1.03 µs	184 ns	788
emit: delegated (v3 selector vs v4 predicate)	v3	1.46 µs	1.52 µs	2.69 µs	1.42 µs	107
emit: delegated (v3 selector vs v4 predicate)	gpu	191 ns	209 ns	893 ns	186 ns	775
emit: 10 core listeners	v3	1.92 µs	1.93 µs	2.09 µs	1.88 µs	80
emit: 10 core listeners	gpu	304 ns	306 ns	365 ns	297 ns	500
emit: on() + off() round-trip	v3	462 ns	549 ns	1.45 µs	427 ns	290
emit: on() + off() round-trip	gpu	215 ns	258 ns	886 ns	182 ns	638
emit: flat vs phased (compound bubbling, 14.5)	orphan (flat emit)	297 ns	378 ns	2.88 µs	276 ns	905069
emit: flat vs phased (compound bubbling, 14.5)	child of 2 ancestors (phased)	834 ns	858 ns	1.30 µs	812 ns	178
emit: flat vs phased, no listeners	orphan (flat)	80.2 ns	87.2 ns	150 ns	78.9 ns	1883
emit: flat vs phased, no listeners	child (phased)	487 ns	528 ns	1.66 µs	475 ns	315
anim: start + stop, one element	v3	923 ns	1.28 µs	3.38 µs	850 ns	402206
anim: start + stop, one element	gpu	634 ns	1.05 µs	4.60 µs	590 ns	464452
anim: start against a running animation (round-21 eviction)	disjoint channels (composes)	1.35 µs	1.93 µs	3.84 µs	1.15 µs	78
anim: start against a running animation (round-21 eviction)	overlapping channels (evicts)	839 ns	1.46 µs	3.61 µs	822 ns	106
anim: start + stop, 512-node collection	v3	296 µs	422 µs	1.02 ms	288 µs	1511
anim: start + stop, 512-node collection	gpu	98.4 µs	112 µs	273 µs	95.2 µs	5729
anim: delay() (no channels touched)	v3	637 ns	934 ns	3.20 µs	619 ns	173
anim: delay() (no channels touched)	gpu	408 ns	708 ns	2.67 µs	403 ns	226
anim: controls (v4 only)	pause + resume	449 ns	732 ns	2.57 µs	444 ns	210
anim: controls (v4 only)	reverse	849 ns	1.47 µs	3.42 µs	837 ns	104
anim: viewport animate + stop	v3	545 ns	809 ns	3.01 µs	532 ns	196
anim: viewport animate + stop	gpu	193 ns	224 ns	1.43 µs	191 ns	739

store N=2,000 · 20.9 s

id index: set x 2000 (build) — gpu

285 µs

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

57.3 ns

id index: single-key operations — has

51.0 ns

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

8.97 ns

id index: single-key operations — hashAt

6.97 ns

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

235 ns

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

56.8 µs

adjacency: reads — outDegree (O(1))

6.24 ns

adjacency: reads — outEdges

45.4 ns

adjacency: reads — connectedEdges

119 ns

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

137 ns

blob: write x 2000 (build) — gpu

105 µs

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

29.9 ns

blob: single-record operations — offsetOf

4.24 ns

blob: single-record operations — free + rewrite

153 ns

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

12.8 ns

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

17.7 ns

dirty: take (a frame's drain) — gpu

590 ns

images: registry bookkeeping — acquire (existing url)

204 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)

11.0 ms

charts: data write refreshing chart-values — gpu

2.37 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	285 µs	298 µs	501 µs	271 µs	2354
id index: single-key operations	get (probe)	57.3 ns	60.4 ns	104 ns	56.1 ns	2837
id index: single-key operations	has	51.0 ns	52.8 ns	94.2 ns	49.4 ns	3251
id index: single-key operations	idAt (memoized hit)	8.97 ns	9.30 ns	17.8 ns	8.66 ns	18262
id index: single-key operations	hashAt	6.97 ns	7.25 ns	14.3 ns	6.89 ns	23370
id index: remove + re-set (churn, drives the round-11 reclaim)	gpu	235 ns	248 ns	378 ns	225 ns	674
adjacency: rebuild (4000 edges, two counting passes)	gpu	56.8 µs	64.7 µs	230 µs	54.9 µs	10741
adjacency: reads	outDegree (O(1))	6.24 ns	6.35 ns	10.5 ns	5.83 ns	26713
adjacency: reads	outEdges	45.4 ns	46.0 ns	60.0 ns	43.0 ns	3715
adjacency: reads	connectedEdges	119 ns	181 ns	2.15 µs	114 ns	923
adjacency: addEdge + removeEdge (the overlay path)	gpu	137 ns	140 ns	208 ns	130 ns	1218
blob: write x 2000 (build)	gpu	105 µs	127 µs	647 µs	101 µs	5500
blob: single-record operations	rewrite in place (same length)	29.9 ns	30.6 ns	43.1 ns	28.9 ns	5604
blob: single-record operations	offsetOf	4.24 ns	4.36 ns	7.92 ns	4.10 ns	38555
blob: single-record operations	free + rewrite	153 ns	174 ns	494 ns	148 ns	981
dirty: mark (per-write span coalescing)	contiguous marks	12.8 ns	13.2 ns	24.5 ns	12.3 ns	12851
dirty: mark (per-write span coalescing)	scattered marks	17.7 ns	18.2 ns	28.6 ns	17.1 ns	9379
dirty: take (a frame's drain)	gpu	590 ns	694 ns	1.33 µs	582 ns	237
images: registry bookkeeping	acquire (existing url)	204 ns	211 ns	271 ns	203 ns	803
images: registry bookkeeping	acquire (new url) + release	534 ns	602 ns	2.64 µs	522 ns	273
charts: sheet apply, chart vs no chart	no chart	10.4 ms	10.5 ms	13.1 ms	10.1 ms	61
charts: sheet apply, chart vs no chart	pie from data (blob record per node)	11.0 ms	11.1 ms	12.0 ms	10.5 ms	59
charts: data write refreshing chart-values	gpu	2.37 ms	2.44 ms	3.54 ms	2.32 ms	282

surface N=2,000 · 2.9 minv4 (gpu) v3

viewport: zoom() get

4.6×

viewport: zoom( n ) set

1.99×

viewport: pan() get

1.08×

viewport: pan( p ) set

1.84×

viewport: panBy()

1.70×

viewport: extent()

1.52×

viewport: renderedExtent()

1.65×

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

3.9×

viewport: minZoom/maxZoom get (2 calls)

15×

viewport: center( eles )

19×

viewport: reset()

109×

viewport: viewport({ zoom, pan })

61×

viewport: getCenterPan( eles ) — gpu

6.31 µs

core: instanceString + headless + styleEnabled (3 calls)

1.27×

core: hasElementWithId + hasCompoundNodes (2 calls)

1.36×

core: options()

84×

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

1.36×

core: gating getters (5 calls)

1.14×

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

1.87×

core: interaction tuning getters (5 calls) — gpu

7.30 ns

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

8273×

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

17043×

core: batch( fn ) (empty)

6.5×

core: startBatch + endBatch + batching

5.8×

core: collection()

7.3×

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

16.9 µs

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

1.75×

collection: length + empty + nonempty (3 calls)

4.6×

collection: eq/first/last (3 calls)

1.85×

collection: slice()

2.2×

collection: toArray()

12×

collection: sort( fn )

1.60×

collection: reduce( fn )

1.40×

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

1.96×

collection: some + every (2 calls)

1.61×

collection: same + anySame + contains (3 calls)

75×

collection: allAre (all match: the full walk)

1.47×

collection: is (none match: the full walk)

2.5×

collection: union

6.3×

collection: intersection

28×

collection: difference

28×

collection: symmetricDifference

3.8×

collection: absoluteComplement

13×

collection: byGroup

2.7×

collection: diff

3.2×

collection: indexOf

1.29×

traversal: openNeighborhood + closedNeighborhood (2 calls)

2.8×

traversal: connectedNodes

3.3×

traversal: sources + targets (2 calls)

3.3×

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

1.27×

traversal: edgesWith

1.25×

traversal: edgesTo

1.24×

traversal: parallelEdges + codirectedEdges (2 calls)

3.1×

traversal: roots + leaves (2 calls)

15×

traversal: allAreNeighbors

5.0×

traversal: isLoop + isSimple (2 calls)

2.5×

degree: indegree + outdegree (2 calls)

70×

degree: minDegree + maxDegree (2 calls)

40×

degree: minIndegree + maxOutdegree (2 calls)

21×

dimensions: renderedPosition get

2.2×

dimensions: relativePosition get

11×

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

2.6×

dimensions: rendered dims (4 calls)

2.9×

dimensions: renderedBoundingBox

1.86×

dimensions: midpoint + renderedMidpoint (2 calls) — gpu

1.22 µs

dimensions: source/targetEndpoint (2 calls) — gpu

405 ns

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

421 ns

dimensions: shift + silentShift (round-trip)

62×

dimensions: silentPosition

15×

dimensions: positions( fn ) over 100

57×

dimensions: padding + paddedWidth + paddedHeight (3 calls)

13×

curves: isBundledBezier — gpu

24.9 ns

curves: controlPoints + renderedControlPoints (2 calls) — gpu

545 ns

curves: segmentPoints (undefined for bezier) — gpu

41.9 ns

compound: parent + parents + ancestors (3 calls)

3.5×

compound: children + descendants + siblings (3 calls)

4.9×

compound: orphans + nonorphans (2 calls)

7.0×

compound: commonAncestors

128×

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

1.56×

flags: select + unselect + selected (band)

16×

flags: selectify + unselectify + selectable

96×

flags: lock + unlock + locked

94×

flags: grabify + ungrabify + grabbable

96×

flags: show + hide + visible

208×

flags: active/activate/unactivate (3 calls)

15×

flags: panify + unpanify + pannable

97×

flags: takesUpSpace + interactive + transparent (3 calls)

2.7×

flags: effectiveOpacity

2.6×

element data: scratch get + set

7.8×

element data: json() one element

2.2×

element data: jsons() over 100

1.75×

element data: label() — gpu

15.6 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()

2.5×

core: onRender + offRender

2.3×

core: removeScratch()

9536×

core: selectionType get

1.99×

core: isReady()

2.2×

core: window()

2.8×

core: container()

2.6×

core: destroyed()

2.2×

core: removeAllListeners()

8.7×

collection: cy()

3.6×

collection: indexOfId()

1.63×

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

1.17×

collection: silentPositions( fn ) over 100

34×

collection: animated()

2.9×

collection: delayAnimation()

2.8×

collection: removeScratch()

9.6×

collection: component()

42×

collection: layoutDimensions()

3.1×

degree: maxIndegree + minOutdegree (2 calls)

24×

flags: grabbed + hidden + inactive (3 calls)

2.1×

curves: renderedSegmentPoints — gpu

36.7 ns

layout contract: edgeSlots + endpoints + degreeOf — gpu

35.3 µs

formats: isColumnarElements + isSerializedElements + isPackedIds — gpu

4.71 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	5.58 ns	5.72 ns	11.1 ns	5.39 ns	29442
viewport: zoom() get	gpu	1.21 ns	1.24 ns	2.24 ns	1.18 ns	128276
viewport: zoom( n ) set	v3	639 ns	644 ns	696 ns	615 ns	257
viewport: zoom( n ) set	gpu	321 ns	347 ns	789 ns	312 ns	480
viewport: pan() get	v3	5.27 ns	5.41 ns	10.4 ns	5.06 ns	31157
viewport: pan() get	gpu	4.86 ns	5.07 ns	12.4 ns	4.75 ns	33253
viewport: pan( p ) set	v3	635 ns	660 ns	1.19 µs	618 ns	254
viewport: pan( p ) set	gpu	345 ns	355 ns	536 ns	324 ns	466
viewport: panBy()	v3	602 ns	609 ns	684 ns	588 ns	273
viewport: panBy()	gpu	354 ns	365 ns	723 ns	322 ns	461
viewport: extent()	v3	29.6 ns	44.2 ns	201 ns	28.3 ns	3832
viewport: extent()	gpu	19.4 ns	26.3 ns	109 ns	18.2 ns	6377
viewport: renderedExtent()	v3	27.0 ns	28.5 ns	50.7 ns	25.6 ns	5943
viewport: renderedExtent()	gpu	16.3 ns	22.2 ns	98.4 ns	15.1 ns	7588
viewport: width() + height() + size() (3 calls)	v3	21.7 ns	22.7 ns	45.4 ns	20.8 ns	7461
viewport: width() + height() + size() (3 calls)	gpu	5.55 ns	5.87 ns	12.8 ns	5.30 ns	28635
viewport: minZoom/maxZoom get (2 calls)	v3	14.6 ns	15.2 ns	25.2 ns	14.0 ns	11207
viewport: minZoom/maxZoom get (2 calls)	gpu	0.999 ns	1.03 ns	1.42 ns	0.977 ns	151707
viewport: center( eles )	v3	117 µs	122 µs	191 µs	113 µs	5729
viewport: center( eles )	gpu	6.20 µs	6.20 µs	6.29 µs	6.10 µs	23
viewport: reset()	v3	974 ns	1.07 µs	1.65 µs	946 ns	156
viewport: reset()	gpu	8.91 ns	9.26 ns	15.9 ns	8.56 ns	18224
viewport: viewport({ zoom, pan })	v3	970 ns	1.06 µs	1.61 µs	948 ns	157
viewport: viewport({ zoom, pan })	gpu	15.8 ns	16.8 ns	37.3 ns	14.9 ns	10108
viewport: getCenterPan( eles )	gpu	6.31 µs	6.35 µs	6.59 µs	6.22 µs	23
core: instanceString + headless + styleEnabled (3 calls)	v3	32.8 ns	34.2 ns	68.4 ns	31.5 ns	4983
core: instanceString + headless + styleEnabled (3 calls)	gpu	25.8 ns	27.2 ns	53.4 ns	24.9 ns	6268
core: hasElementWithId + hasCompoundNodes (2 calls)	v3	42.7 ns	44.2 ns	85.5 ns	42.6 ns	3879
core: hasElementWithId + hasCompoundNodes (2 calls)	gpu	31.5 ns	33.5 ns	67.8 ns	30.7 ns	5097
core: options()	v3	89.8 ns	93.9 ns	191 ns	86.7 ns	1818
core: options()	gpu	1.07 ns	1.18 ns	2.24 ns	1.03 ns	134854
core: mutableElements() (v4 materializes; v3 is O(1))	v3	3.19 ns	3.27 ns	6.49 ns	3.07 ns	50177
core: mutableElements() (v4 materializes; v3 is O(1))	gpu	2.35 ns	2.40 ns	3.44 ns	2.20 ns	69028
core: gating getters (5 calls)	v3	15.6 ns	16.2 ns	25.8 ns	15.1 ns	10583
core: gating getters (5 calls)	gpu	13.8 ns	14.5 ns	34.4 ns	12.5 ns	11787
core: gating setters round-trip (autolock/autoungrabify/autounselectify)	v3	11.2 ns	11.5 ns	18.6 ns	10.5 ns	14849
core: gating setters round-trip (autolock/autoungrabify/autounselectify)	gpu	6.00 ns	6.18 ns	13.3 ns	5.74 ns	27342
core: interaction tuning getters (5 calls)	gpu	7.30 ns	7.60 ns	16.0 ns	7.00 ns	22356
core: graph data() set + get (v3 restyles the graph)	v3	2.81 ms	2.87 ms	3.79 ms	2.68 ms	239
core: graph data() set + get (v3 restyles the graph)	gpu	340 ns	347 ns	470 ns	338 ns	488
core: graph scratch() set + get (v3 restyles the graph)	v3	2.83 ms	2.93 ms	3.90 ms	2.79 ms	235
core: graph scratch() set + get (v3 restyles the graph)	gpu	166 ns	170 ns	226 ns	162 ns	1006
core: batch( fn ) (empty)	v3	542 ns	646 ns	1.51 µs	521 ns	260
core: batch( fn ) (empty)	gpu	82.9 ns	86.9 ns	143 ns	81.5 ns	1953
core: startBatch + endBatch + batching	v3	523 ns	628 ns	1.63 µs	509 ns	268
core: startBatch + endBatch + batching	gpu	89.7 ns	165 ns	2.90 µs	86.5 ns	998
core: collection()	v3	452 ns	531 ns	1.51 µs	445 ns	317
core: collection()	gpu	62.0 ns	65.4 ns	108 ns	60.4 ns	2597
core: compact() (no dead slots — the early-out)	gpu	16.9 µs	21.8 µs	99.1 µs	16.4 µs	31360
collection: element accessors (id/group/isNode/isEdge, 4 calls)	v3	60.6 ns	64.4 ns	196 ns	56.4 ns	2644
collection: element accessors (id/group/isNode/isEdge, 4 calls)	gpu	34.6 ns	36.1 ns	68.6 ns	33.6 ns	4722
collection: length + empty + nonempty (3 calls)	v3	6.78 ns	6.98 ns	13.1 ns	6.47 ns	24270
collection: length + empty + nonempty (3 calls)	gpu	1.47 ns	1.52 ns	1.90 ns	1.44 ns	107177
collection: eq/first/last (3 calls)	v3	15.3 ns	15.6 ns	23.5 ns	14.8 ns	10946
collection: eq/first/last (3 calls)	gpu	8.24 ns	8.46 ns	14.7 ns	7.93 ns	20072
collection: slice()	v3	3.81 µs	3.92 µs	4.29 µs	3.69 µs	39
collection: slice()	gpu	1.73 µs	1.74 µs	1.96 µs	1.67 µs	93
collection: toArray()	v3	8.85 µs	8.96 µs	9.06 µs	8.63 µs	15
collection: toArray()	gpu	759 ns	774 ns	903 ns	734 ns	211
collection: sort( fn )	v3	14.5 µs	14.9 µs	15.1 µs	14.0 µs	12
collection: sort( fn )	gpu	9.08 µs	9.10 µs	9.17 µs	9.02 µs	15
collection: reduce( fn )	v3	157 ns	160 ns	194 ns	120 ns	1065
collection: reduce( fn )	gpu	112 ns	114 ns	133 ns	109 ns	1500
collection: min + max ( fn ) (2 calls)	v3	11.0 µs	11.1 µs	11.4 µs	10.9 µs	12
collection: min + max ( fn ) (2 calls)	gpu	5.64 µs	5.73 µs	6.59 µs	5.57 µs	25
collection: some + every (2 calls)	v3	894 ns	922 ns	1.58 µs	871 ns	180
collection: some + every (2 calls)	gpu	557 ns	560 ns	613 ns	539 ns	303
collection: same + anySame + contains (3 calls)	v3	7.50 µs	7.53 µs	7.76 µs	7.40 µs	19
collection: same + anySame + contains (3 calls)	gpu	99.5 ns	102 ns	128 ns	95.5 ns	1686
collection: allAre (all match: the full walk)	v3	3.63 µs	3.62 µs	3.96 µs	3.35 µs	43
collection: allAre (all match: the full walk)	gpu	2.48 µs	2.48 µs	2.84 µs	2.34 µs	64
collection: is (none match: the full walk)	v3	5.73 µs	5.75 µs	6.17 µs	5.45 µs	26
collection: is (none match: the full walk)	gpu	2.29 µs	2.33 µs	2.79 µs	2.23 µs	67
collection: union	v3	95.4 µs	105 µs	184 µs	93.4 µs	6601
collection: union	gpu	15.1 µs	15.3 µs	15.7 µs	14.7 µs	12
collection: intersection	v3	92.5 µs	103 µs	164 µs	90.2 µs	6717
collection: intersection	gpu	3.31 µs	3.34 µs	3.65 µs	3.24 µs	46
collection: difference	v3	92.6 µs	103 µs	168 µs	88.8 µs	6714
collection: difference	gpu	3.27 µs	3.28 µs	3.44 µs	3.18 µs	47
collection: symmetricDifference	v3	25.4 µs	25.9 µs	27.0 µs	24.7 µs	12
collection: symmetricDifference	gpu	6.76 µs	6.75 µs	6.83 µs	6.64 µs	21
collection: absoluteComplement	v3	6.19 ms	6.77 ms	10.6 ms	5.95 ms	99
collection: absoluteComplement	gpu	462 µs	475 µs	769 µs	452 µs	1473
collection: byGroup	v3	998 µs	1.02 ms	1.52 ms	972 µs	682
collection: byGroup	gpu	375 µs	386 µs	626 µs	365 µs	1810
collection: diff	v3	37.6 µs	38.4 µs	41.4 µs	35.7 µs	12
collection: diff	gpu	11.8 µs	11.7 µs	12.2 µs	10.2 µs	11
collection: indexOf	v3	29.9 ns	31.1 ns	52.1 ns	28.5 ns	5484
collection: indexOf	gpu	23.2 ns	24.0 ns	38.3 ns	22.1 ns	7131
traversal: openNeighborhood + closedNeighborhood (2 calls)	v3	9.24 µs	9.42 µs	9.74 µs	8.74 µs	14
traversal: openNeighborhood + closedNeighborhood (2 calls)	gpu	3.25 µs	3.33 µs	3.72 µs	3.13 µs	47
traversal: connectedNodes	v3	979 ns	1.06 µs	1.84 µs	925 ns	156
traversal: connectedNodes	gpu	298 ns	306 ns	401 ns	288 ns	542
traversal: sources + targets (2 calls)	v3	1.73 µs	1.90 µs	2.78 µs	1.67 µs	86
traversal: sources + targets (2 calls)	gpu	525 ns	680 ns	2.13 µs	501 ns	242
traversal: source() + target() (2 calls)	v3	72.6 ns	74.6 ns	101 ns	60.8 ns	2304
traversal: source() + target() (2 calls)	gpu	57.0 ns	59.1 ns	79.5 ns	54.5 ns	2895
traversal: edgesWith	v3	71.2 µs	73.2 µs	112 µs	68.8 µs	9518
traversal: edgesWith	gpu	57.1 µs	62.2 µs	120 µs	55.4 µs	11117
traversal: edgesTo	v3	72.7 µs	74.8 µs	128 µs	71.5 µs	9322
traversal: edgesTo	gpu	58.7 µs	59.5 µs	62.9 µs	57.4 µs	12
traversal: parallelEdges + codirectedEdges (2 calls)	v3	3.52 µs	3.79 µs	4.55 µs	3.36 µs	41
traversal: parallelEdges + codirectedEdges (2 calls)	gpu	1.15 µs	1.29 µs	1.85 µs	1.12 µs	127
traversal: roots + leaves (2 calls)	v3	4.99 ms	5.35 ms	8.68 ms	4.88 ms	127
traversal: roots + leaves (2 calls)	gpu	342 µs	347 µs	397 µs	337 µs	2024
traversal: allAreNeighbors	v3	195 µs	210 µs	312 µs	189 µs	3312
traversal: allAreNeighbors	gpu	38.7 µs	39.7 µs	41.9 µs	38.3 µs	12
traversal: isLoop + isSimple (2 calls)	v3	127 ns	129 ns	148 ns	123 ns	1326
traversal: isLoop + isSimple (2 calls)	gpu	51.2 ns	52.4 ns	65.1 ns	50.3 ns	3265
degree: indegree + outdegree (2 calls)	v3	3.52 µs	3.77 µs	4.53 µs	3.40 µs	41
degree: indegree + outdegree (2 calls)	gpu	50.1 ns	53.9 ns	110 ns	49.3 ns	3171
degree: minDegree + maxDegree (2 calls)	v3	17.9 ms	17.8 ms	21.9 ms	15.5 ms	35
degree: minDegree + maxDegree (2 calls)	gpu	443 µs	457 µs	671 µs	436 µs	1530
degree: minIndegree + maxOutdegree (2 calls)	v3	9.16 ms	9.89 ms	13.4 ms	8.85 ms	67
degree: minIndegree + maxOutdegree (2 calls)	gpu	426 µs	448 µs	872 µs	410 µs	1562
dimensions: renderedPosition get	v3	67.2 ns	69.6 ns	115 ns	64.1 ns	2424
dimensions: renderedPosition get	gpu	29.9 ns	33.4 ns	100 ns	28.8 ns	5040
dimensions: relativePosition get	v3	838 ns	847 ns	975 ns	816 ns	197
dimensions: relativePosition get	gpu	73.6 ns	78.5 ns	138 ns	71.0 ns	2154
dimensions: width + height + outerWidth + outerHeight (4 calls)	v3	604 ns	608 ns	661 ns	590 ns	278
dimensions: width + height + outerWidth + outerHeight (4 calls)	gpu	232 ns	239 ns	312 ns	228 ns	712
dimensions: rendered dims (4 calls)	v3	700 ns	705 ns	794 ns	669 ns	240
dimensions: rendered dims (4 calls)	gpu	245 ns	248 ns	335 ns	234 ns	690
dimensions: renderedBoundingBox	v3	1.73 µs	1.76 µs	2.25 µs	1.66 µs	93
dimensions: renderedBoundingBox	gpu	933 ns	953 ns	1.33 µs	894 ns	175
dimensions: midpoint + renderedMidpoint (2 calls)	gpu	1.22 µs	1.24 µs	1.43 µs	1.19 µs	130
dimensions: source/targetEndpoint (2 calls)	gpu	405 ns	415 ns	518 ns	397 ns	409
dimensions: rendered source/targetEndpoint (2 calls)	gpu	421 ns	431 ns	572 ns	407 ns	394
dimensions: shift + silentShift (round-trip)	v3	293 µs	311 µs	591 µs	282 µs	2242
dimensions: shift + silentShift (round-trip)	gpu	4.77 µs	4.79 µs	4.89 µs	4.71 µs	32
dimensions: silentPosition	v3	868 ns	871 ns	932 ns	847 ns	192
dimensions: silentPosition	gpu	56.6 ns	58.0 ns	84.1 ns	54.2 ns	2919
dimensions: positions( fn ) over 100	v3	175 µs	180 µs	255 µs	172 µs	3876
dimensions: positions( fn ) over 100	gpu	3.05 µs	3.12 µs	3.34 µs	2.99 µs	51
dimensions: padding + paddedWidth + paddedHeight (3 calls)	v3	2.02 µs	2.04 µs	2.31 µs	1.98 µs	80
dimensions: padding + paddedWidth + paddedHeight (3 calls)	gpu	160 ns	164 ns	210 ns	157 ns	1039
curves: isBundledBezier	gpu	24.9 ns	25.7 ns	38.5 ns	21.2 ns	6618
curves: controlPoints + renderedControlPoints (2 calls)	gpu	545 ns	555 ns	729 ns	533 ns	298
curves: segmentPoints (undefined for bezier)	gpu	41.9 ns	42.7 ns	53.5 ns	35.7 ns	4002
compound: parent + parents + ancestors (3 calls)	v3	2.89 µs	3.26 µs	4.13 µs	2.81 µs	48
compound: parent + parents + ancestors (3 calls)	gpu	821 ns	1.07 µs	2.13 µs	779 ns	153
compound: children + descendants + siblings (3 calls)	v3	9.05 µs	9.14 µs	9.88 µs	8.35 µs	15
compound: children + descendants + siblings (3 calls)	gpu	1.83 µs	2.24 µs	3.44 µs	1.75 µs	72
compound: orphans + nonorphans (2 calls)	v3	3.71 µs	4.01 µs	4.81 µs	3.58 µs	38
compound: orphans + nonorphans (2 calls)	gpu	531 ns	688 ns	2.64 µs	517 ns	238
compound: commonAncestors	v3	9.50 µs	9.62 µs	10.1 µs	9.10 µs	14
compound: commonAncestors	gpu	74.5 ns	76.8 ns	117 ns	72.0 ns	2175
compound: isParent/isChild/isChildless/isOrphan (4 calls)	v3	70.5 ns	73.1 ns	136 ns	60.6 ns	2338
compound: isParent/isChild/isChildless/isOrphan (4 calls)	gpu	45.2 ns	47.0 ns	74.3 ns	44.0 ns	3643
flags: select + unselect + selected (band)	v3	197 µs	204 µs	367 µs	192 µs	3423
flags: select + unselect + selected (band)	gpu	12.3 µs	12.3 µs	12.5 µs	12.2 µs	10
flags: selectify + unselectify + selectable	v3	197 µs	203 µs	315 µs	192 µs	3443
flags: selectify + unselectify + selectable	gpu	2.06 µs	2.07 µs	2.16 µs	2.01 µs	79
flags: lock + unlock + locked	v3	195 µs	204 µs	440 µs	188 µs	3426
flags: lock + unlock + locked	gpu	2.06 µs	2.07 µs	2.26 µs	1.99 µs	79
flags: grabify + ungrabify + grabbable	v3	199 µs	210 µs	449 µs	193 µs	3322
flags: grabify + ungrabify + grabbable	gpu	2.08 µs	2.10 µs	2.33 µs	2.01 µs	78
flags: show + hide + visible	v3	2.61 ms	2.83 ms	4.64 ms	2.54 ms	243
flags: show + hide + visible	gpu	12.5 µs	12.8 µs	13.3 µs	12.4 µs	9
flags: active/activate/unactivate (3 calls)	v3	200 µs	210 µs	396 µs	194 µs	3330
flags: active/activate/unactivate (3 calls)	gpu	13.3 µs	13.4 µs	13.7 µs	13.2 µs	9
flags: panify + unpanify + pannable	v3	199 µs	211 µs	400 µs	192 µs	3309
flags: panify + unpanify + pannable	gpu	2.05 µs	2.14 µs	2.58 µs	2.04 µs	76
flags: takesUpSpace + interactive + transparent (3 calls)	v3	199 ns	207 ns	354 ns	198 ns	825
flags: takesUpSpace + interactive + transparent (3 calls)	gpu	72.9 ns	77.4 ns	143 ns	71.4 ns	2208
flags: effectiveOpacity	v3	55.2 ns	57.7 ns	129 ns	52.8 ns	2970
flags: effectiveOpacity	gpu	21.6 ns	23.6 ns	47.6 ns	21.0 ns	7237
element data: scratch get + set	v3	1.37 µs	1.50 µs	3.08 µs	1.32 µs	109
element data: scratch get + set	gpu	175 ns	194 ns	339 ns	174 ns	876
element data: json() one element	v3	311 ns	428 ns	3.70 µs	296 ns	380
element data: json() one element	gpu	141 ns	189 ns	535 ns	139 ns	846
element data: jsons() over 100	v3	43.1 µs	43.6 µs	46.9 µs	39.5 µs	12
element data: jsons() over 100	gpu	24.6 µs	23.9 µs	29.8 µs	16.0 µs	12
element data: label()	gpu	15.6 ns	17.8 ns	34.0 ns	14.9 ns	9619
core: batchData()	v3	2.43 µs	2.55 µs	2.97 µs	2.36 µs	62
core: batchData()	gpu	282 ns	376 ns	2.22 µs	280 ns	438
core: animated()	v3	7.69 ns	8.05 ns	15.1 ns	7.52 ns	21048
core: animated()	gpu	1.56 ns	1.62 ns	2.65 ns	1.56 ns	101003
core: zoomRange get	gpu	2.16 ns	2.36 ns	6.30 ns	2.05 ns	70472
core: forceRender() (headless no-op)	v3	23.8 ns	24.5 ns	42.0 ns	23.5 ns	6993
core: forceRender() (headless no-op)	gpu	1.51 ns	1.56 ns	2.35 ns	1.51 ns	105153
core: resize() / invalidateSize()	v3	333 ns	345 ns	473 ns	330 ns	492
core: resize() / invalidateSize()	gpu	135 ns	142 ns	211 ns	133 ns	1196
core: onRender + offRender	v3	451 ns	476 ns	779 ns	446 ns	352
core: onRender + offRender	gpu	193 ns	238 ns	730 ns	190 ns	692
core: removeScratch()	v3	2.82 ms	2.94 ms	4.07 ms	2.75 ms	233
core: removeScratch()	gpu	296 ns	311 ns	646 ns	277 ns	546
core: selectionType get	v3	4.60 ns	4.96 ns	9.93 ns	4.44 ns	34109
core: selectionType get	gpu	2.31 ns	2.37 ns	3.62 ns	2.15 ns	70216
core: isReady()	v3	3.32 ns	3.47 ns	6.34 ns	3.32 ns	48405
core: isReady()	gpu	1.51 ns	1.55 ns	2.07 ns	1.48 ns	105607
core: window()	v3	4.19 ns	4.32 ns	6.18 ns	4.18 ns	38914
core: window()	gpu	1.51 ns	1.56 ns	2.32 ns	1.51 ns	105224
core: container()	v3	3.96 ns	4.10 ns	6.12 ns	3.82 ns	41065
core: container()	gpu	1.51 ns	1.56 ns	2.36 ns	1.51 ns	105230
core: destroyed()	v3	3.33 ns	3.43 ns	4.61 ns	3.32 ns	48877
core: destroyed()	gpu	1.51 ns	1.56 ns	2.66 ns	1.48 ns	105018
core: removeAllListeners()	v3	14.1 ns	15.6 ns	56.8 ns	13.5 ns	10929
core: removeAllListeners()	gpu	1.62 ns	1.71 ns	3.15 ns	1.59 ns	95713
collection: cy()	v3	7.85 ns	9.42 ns	25.0 ns	7.48 ns	18071
collection: cy()	gpu	2.20 ns	2.58 ns	6.33 ns	2.15 ns	64625
collection: indexOfId()	v3	60.3 ns	62.8 ns	106 ns	59.3 ns	2720
collection: indexOfId()	gpu	36.9 ns	38.0 ns	57.3 ns	35.6 ns	4473
collection: removed() + inside() (2 calls)	v3	18.1 ns	19.3 ns	34.4 ns	17.4 ns	8835
collection: removed() + inside() (2 calls)	gpu	15.4 ns	16.2 ns	32.4 ns	15.2 ns	10523
collection: silentPositions( fn ) over 100	v3	110 µs	118 µs	264 µs	107 µs	5887
collection: silentPositions( fn ) over 100	gpu	3.22 µs	3.41 µs	4.51 µs	3.07 µs	45
collection: animated()	v3	12.4 ns	12.6 ns	16.9 ns	12.2 ns	13483
collection: animated()	gpu	4.30 ns	4.80 ns	10.6 ns	4.17 ns	35201
collection: delayAnimation()	v3	453 ns	607 ns	2.07 µs	450 ns	273
collection: delayAnimation()	gpu	164 ns	171 ns	302 ns	161 ns	981
collection: removeScratch()	v3	1.97 µs	2.05 µs	3.34 µs	1.88 µs	77
collection: removeScratch()	gpu	205 ns	212 ns	346 ns	200 ns	804
collection: component()	v3	83.8 ms	86.9 ms	98.7 ms	78.1 ms	12
collection: component()	gpu	2.02 ms	2.18 ms	3.67 ms	1.98 ms	310
collection: layoutDimensions()	v3	621 ns	636 ns	836 ns	619 ns	264
collection: layoutDimensions()	gpu	203 ns	212 ns	351 ns	195 ns	801
degree: maxIndegree + minOutdegree (2 calls)	v3	10.5 ms	11.0 ms	15.9 ms	9.26 ms	60
degree: maxIndegree + minOutdegree (2 calls)	gpu	436 µs	481 µs	853 µs	421 µs	1455
flags: grabbed + hidden + inactive (3 calls)	v3	87.3 ns	89.9 ns	129 ns	80.9 ns	1902
flags: grabbed + hidden + inactive (3 calls)	gpu	42.2 ns	43.5 ns	64.8 ns	40.5 ns	3941
curves: renderedSegmentPoints	gpu	36.7 ns	38.2 ns	83.8 ns	36.0 ns	4479
layout contract: edgeSlots + endpoints + degreeOf	gpu	35.3 µs	40.0 µs	67.0 µs	30.2 µs	17186
formats: isColumnarElements + isSerializedElements + isPackedIds	gpu	4.71 ns	5.39 ns	9.39 ns	4.70 ns	31406

compaction N=20,000 · 19.0 s

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

227 µs

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

51.4 µs

cpu pick: node hit — peak

226 µs

cpu pick: node hit — compacted

51.2 µs

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

6.08 ns

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

4.60 ns

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

9.17 ns

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

46.2 ns

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

23.6 µs

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

23.4 µs

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

117 µs

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

114 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	227 µs	259 µs	478 µs	226 µs	2710
cpu pick: background miss (full highWater walk)	compacted	51.4 µs	51.9 µs	53.3 µs	50.8 µs	12
cpu pick: node hit	peak	226 µs	234 µs	391 µs	225 µs	2995
cpu pick: node hit	compacted	51.2 µs	52.4 µs	55.5 µs	50.8 µs	12
isCurrent: current ref (forwards present vs absent)	no forwards (peak)	6.08 ns	6.69 ns	14.4 ns	5.86 ns	25233
isCurrent: current ref (forwards present vs absent)	forwards present (compacted)	4.60 ns	5.24 ns	11.9 ns	4.53 ns	31988
isCurrent: stale-ref repair (chase + rewrite, alloc included)	current (baseline alloc)	9.17 ns	12.0 ns	37.5 ns	8.56 ns	13934
isCurrent: stale-ref repair (chase + rewrite, alloc included)	stale → repair	46.2 ns	49.0 ns	106 ns	44.6 ns	3471
control: nodes({ selected }) scan (order-list, expected parity)	peak	23.6 µs	27.4 µs	35.4 µs	22.1 µs	12
control: nodes({ selected }) scan (order-list, expected parity)	compacted	23.4 µs	25.3 µs	30.4 µs	21.5 µs	12
control: whole-graph boundingBox (expected parity)	peak	117 µs	129 µs	263 µs	114 µs	5416
control: whole-graph boundingBox (expected parity)	compacted	114 µs	116 µs	176 µs	111 µs	6019

compound N=2,000 · 9.4 sv4 (gpu) v3

parent drag (shift subtree + bb settle)

186×

child drag (move + parent re-derive)

51×

reparent round-trip

74×

1 µs10 µs100 µs1 ms

All stats (table view)

group	bench	p50	avg	p99	min	samples
parent drag (shift subtree + bb settle)	v3	267 µs	305 µs	817 µs	254 µs	2280
parent drag (shift subtree + bb settle)	gpu	1.44 µs	1.45 µs	1.62 µs	1.38 µs	108
child drag (move + parent re-derive)	v3	83.1 µs	91.6 µs	163 µs	79.2 µs	7470
child drag (move + parent re-derive)	gpu	1.62 µs	1.66 µs	2.02 µs	1.59 µs	96
reparent round-trip	v3	84.9 µs	104 µs	249 µs	81.6 µs	6628
reparent round-trip	gpu	1.15 µs	1.18 µs	1.56 µs	1.10 µs	135

curves N=2,000 · 0.7 s

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

31.8 ms

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

28.4 ms

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

536 µs

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

314 µs

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

368 µs

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

399 µs

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

9.63 ms

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

3.22 ms

curve premium: same read again (warm) — curved

2.85 ms

curve premium: same read again (warm) — straight

1.75 ms

curve premium: edge.midpoint x 4000 — curved

4.43 ms

curve premium: edge.midpoint x 4000 — straight

3.36 ms

curve premium: edge.controlPoints x 4000 — curved

2.30 ms

curve premium: edge.controlPoints x 4000 — straight

1.36 ms

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

36.3 ms

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

22.7 ms

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

4.00 ms

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

3.73 ms

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

27.7 ms

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

7.24 ms

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

4.23 ms

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

2.39 ms

100 µs1 ms10 ms100 ms

All stats (table view)

group	bench	p50	avg	p99	min	samples
curve premium: build (style apply + derivation)	curved	31.8 ms	31.8 ms	31.8 ms	31.8 ms	1
curve premium: build (style apply + derivation)	straight	28.4 ms	28.4 ms	28.4 ms	28.4 ms	1
curve premium: node.position x 1000 (drag)	curved	536 µs	536 µs	536 µs	536 µs	1
curve premium: node.position x 1000 (drag)	straight	314 µs	314 µs	314 µs	314 µs	1
curve premium: nodes().positions() (bulk move)	curved	368 µs	368 µs	368 µs	368 µs	1
curve premium: nodes().positions() (bulk move)	straight	399 µs	399 µs	399 µs	399 µs	1
curve premium: first read after a bulk move (cold)	curved	9.63 ms	9.63 ms	9.63 ms	9.63 ms	1
curve premium: first read after a bulk move (cold)	straight	3.22 ms	3.22 ms	3.22 ms	3.22 ms	1
curve premium: same read again (warm)	curved	2.85 ms	2.85 ms	2.85 ms	2.85 ms	1
curve premium: same read again (warm)	straight	1.75 ms	1.75 ms	1.75 ms	1.75 ms	1
curve premium: edge.midpoint x 4000	curved	4.43 ms	4.43 ms	4.43 ms	4.43 ms	1
curve premium: edge.midpoint x 4000	straight	3.36 ms	3.36 ms	3.36 ms	3.36 ms	1
curve premium: edge.controlPoints x 4000	curved	2.30 ms	2.30 ms	2.30 ms	2.30 ms	1
curve premium: edge.controlPoints x 4000	straight	1.36 ms	1.36 ms	1.36 ms	1.36 ms	1
curve premium: cy.elements().boundingBox() x 20	curved	36.3 ms	36.3 ms	36.3 ms	36.3 ms	1
curve premium: cy.elements().boundingBox() x 20	straight	22.7 ms	22.7 ms	22.7 ms	22.7 ms	1
curve premium: cy.fit() x 20 (store scan)	curved	4.00 ms	4.00 ms	4.00 ms	4.00 ms	1
curve premium: cy.fit() x 20 (store scan)	straight	3.73 ms	3.73 ms	3.73 ms	3.73 ms	1
curve premium: elementsInBox x 20 (half the graph)	curved	27.7 ms	27.7 ms	27.7 ms	27.7 ms	1
curve premium: elementsInBox x 20 (half the graph)	straight	7.24 ms	7.24 ms	7.24 ms	7.24 ms	1
curve premium: edge.hide/show x 1000 (bundle re-fan)	curved	4.23 ms	4.23 ms	4.23 ms	4.23 ms	1
curve premium: edge.hide/show x 1000 (bundle re-fan)	straight	2.39 ms	2.39 ms	2.39 ms	2.39 ms	1

arrows N=2,000 · 3.4 s

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

24.5 ms

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

22.2 ms

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

43.6 ms

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

44.0 ms

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

23.5 ms

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

23.4 ms

arrow premium: edge.sourceEndpoint x 200000 — arrowed

60.2 ms

arrow premium: edge.sourceEndpoint x 200000 — none

58.5 ms

arrow premium: edge.targetEndpoint x 200000 — arrowed

60.8 ms

arrow premium: edge.targetEndpoint x 200000 — none

62.3 ms

arrow premium: edge.midpoint x 200000 — arrowed

139 ms

arrow premium: edge.midpoint x 200000 — none

137 ms

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

34.0 ms

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

30.6 ms

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

21.2 ms

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

21.2 ms

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

5.54 ms

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

5.48 ms

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

73.4 ms

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

58.5 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	24.5 ms	24.5 ms	24.5 ms	24.5 ms	1
arrow premium: build (style apply + arrow fold)	none	22.2 ms	22.2 ms	22.2 ms	22.2 ms	1
arrow premium: cy.style() re-apply x 4	arrowed	43.6 ms	43.6 ms	43.6 ms	43.6 ms	1
arrow premium: cy.style() re-apply x 4	none	44.0 ms	44.0 ms	44.0 ms	44.0 ms	1
arrow premium: node.position x 200000 (drag)	arrowed	23.5 ms	23.5 ms	23.5 ms	23.5 ms	1
arrow premium: node.position x 200000 (drag)	none	23.4 ms	23.4 ms	23.4 ms	23.4 ms	1
arrow premium: edge.sourceEndpoint x 200000	arrowed	60.2 ms	60.2 ms	60.2 ms	60.2 ms	1
arrow premium: edge.sourceEndpoint x 200000	none	58.5 ms	58.5 ms	58.5 ms	58.5 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	62.3 ms	62.3 ms	62.3 ms	62.3 ms	1
arrow premium: edge.midpoint x 200000	arrowed	139 ms	139 ms	139 ms	139 ms	1
arrow premium: edge.midpoint x 200000	none	137 ms	137 ms	137 ms	137 ms	1
arrow premium: edge.style('target-arrow-shape') x 200000	arrowed	34.0 ms	34.0 ms	34.0 ms	34.0 ms	1
arrow premium: edge.style('target-arrow-shape') x 200000	none	30.6 ms	30.6 ms	30.6 ms	30.6 ms	1
arrow premium: cy.elements().boundingBox() x 20	arrowed	21.2 ms	21.2 ms	21.2 ms	21.2 ms	1
arrow premium: cy.elements().boundingBox() x 20	none	21.2 ms	21.2 ms	21.2 ms	21.2 ms	1
arrow premium: elementsInBox x 20 (half the graph)	arrowed	5.54 ms	5.54 ms	5.54 ms	5.54 ms	1
arrow premium: elementsInBox x 20 (half the graph)	none	5.48 ms	5.48 ms	5.48 ms	5.48 ms	1
arrow premium: control: sourceEndpoint, bezier vs straight	arrowed	73.4 ms	73.4 ms	73.4 ms	73.4 ms	1
arrow premium: control: sourceEndpoint, bezier vs straight	none	58.5 ms	58.5 ms	58.5 ms	58.5 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

83.5 ms

labels: estimateBlock x 20000 — gpu

92.7 ms

labels: addNode x 20000 — gpu

30.8 ms

labels: setLabel (build) x 20000 — gpu

104 ms

labels: setLabel (rewrite) x 20000 — gpu

102 ms

labels: boundingBox with 20000 label terms — gpu

3.40 ms

1 ms10 ms100 ms1 s

All stats (table view)

group	bench	p50	avg	p99	min	samples
labels: breakLines x 20000	gpu	83.5 ms	83.5 ms	83.5 ms	83.5 ms	1
labels: estimateBlock x 20000	gpu	92.7 ms	92.7 ms	92.7 ms	92.7 ms	1
labels: addNode x 20000	gpu	30.8 ms	30.8 ms	30.8 ms	30.8 ms	1
labels: setLabel (build) x 20000	gpu	104 ms	104 ms	104 ms	104 ms	1
labels: setLabel (rewrite) x 20000	gpu	102 ms	102 ms	102 ms	102 ms	1
labels: boundingBox with 20000 label terms	gpu	3.40 ms	3.40 ms	3.40 ms	3.40 ms	1

transitions N=2,000 · 9.5 s

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

2.31 ms

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

4.80 ms

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

136 µs

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

2.45 µs

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

5.90 µs

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

10.0 ms

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

12.8 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	2.31 ms	2.47 ms	4.55 ms	2.19 ms	277
transition: auto-extent shift, whole-channel re-derive (2000 slots)	transitions on	4.80 ms	5.08 ms	7.47 ms	4.52 ms	132
transition: bulk tween tick (2000 slots, one manager tick)	tick	136 µs	143 µs	254 µs	129 µs	4890
transition: explicit-domain write (O(changed) — one element)	transitions off	2.45 µs	2.54 µs	3.77 µs	2.37 µs	62
transition: explicit-domain write (O(changed) — one element)	transitions on	5.90 µs	5.87 µs	6.76 µs	4.94 µs	25
transition: sheet swap, whole-graph apply (2000 slots diff + tween)	transitions off	10.0 ms	10.0 ms	10.6 ms	9.78 ms	66
transition: sheet swap, whole-graph apply (2000 slots diff + tween)	transitions on	12.8 ms	12.6 ms	14.1 ms	11.5 ms	52

geometry-tween N=2,000 · 11.0 s

size tick: unlabelled (2000 slots) — tick

526 µs

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

548 µs

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

1.18 ms

edge width tick (4000 slots) — bare

243 µs

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

724 µs

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

175 µs

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

360 µs

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

5.84 ms

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

147 µs

100 µs1 ms10 ms

All stats (table view)

group	bench	p50	avg	p99	min	samples
size tick: unlabelled (2000 slots)	tick	526 µs	552 µs	917 µs	509 µs	1247
size tick: labelled, center/center (re-anchor early-out) (2000 slots)	tick	548 µs	570 µs	885 µs	531 µs	1228
size tick: labelled, hanging (sidecar rewrite per tick) (2000 slots)	tick	1.18 ms	1.26 ms	2.13 ms	1.13 ms	534
edge width tick (4000 slots)	bare	243 µs	254 µs	400 µs	235 µs	2748
edge width tick (4000 slots)	with rides (casing + match-line arrows)	724 µs	747 µs	1.11 ms	697 µs	936
padding tick + auto-bounds flush (250 parents × 8 children)	tick + flush	175 µs	189 µs	430 µs	171 µs	3684
font-size tick (2000 labels)	wrap none (dims scale-patch)	360 µs	411 µs	942 µs	347 µs	1682
font-size tick (2000 labels)	wrapped (re-break per tick)	5.84 ms	5.91 ms	6.68 ms	5.40 ms	112
baseline: paint (opacity) tick (2000 slots)	tick	147 µs	153 µs	286 µs	138 µs	4580