Cytoscape.js GPU benchmarks

v4 executor 'cpu' vs 'gpu' · times are per-call p50 · 2026-08-10 19:05 · 1f324088 (v4) · chromium v24.18.0 · algorithms-gpu profile

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
Adapter	amd gcn-4
Browser	chromium 149.0.7827.55
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 cpu p50 ÷ gpu p50, log scale · right of the 1× line = gpu faster

markovClustering @ 1.024k

642×

markovClustering @ 512

300×

kMedoids @ 4.096k

163×

kMedoids @ 2.048k

103×

markovClustering @ 256

72×

fuzzyCMeans @ 65.536k

66×

fuzzyCMeans @ 16.384k

49×

floydWarshall @ 1.024k

32×

fuzzyCMeans @ 4.096k

31×

kMedoids @ 1.024k

23×

kMeans @ 65.536k

23×

betweennessCentrality @ 2.048k

18×

kMeans @ 16.384k

17×

floydWarshall @ 512

11×

betweennessCentrality @ 1.024k

10×

kMeans @ 4.096k

7.9×

affinityPropagation @ 1.024k

5.1×

betweennessCentrality @ 512

3.8×

floydWarshall @ 256

3.4×

affinityPropagation @ 256

2.6×

hierarchicalClustering @ 1.024k

2.1×

hierarchicalClustering @ 2.048k

1.89×

affinityPropagation @ 512

1.86×

hierarchicalClustering @ 4.096k

1.74×

pageRank @ 2.048k

1.69×

pageRank @ 1.024k

1.67×

pageRank @ 512

1.27×

1×10×100×1000×

Scaling speedup vs graph size

algorithms-gpu

group	× @ N=256	× @ N=512	× @ N=1.024k	× @ N=2.048k	× @ N=4.096k	× @ N=16.384k	× @ N=65.536k
markovClustering	72×	300×	642×	–	–	–	–
floydWarshall	3.4×	11×	32×	–	–	–	–
affinityPropagation	2.6×	1.86×	5.1×	–	–	–	–
pageRank	–	1.27×	1.67×	1.69×	–	–	–
betweennessCentrality	–	3.8×	10×	18×	–	–	–
kMedoids	–	–	23×	103×	163×	–	–
hierarchicalClustering	–	–	2.1×	1.89×	1.74×	–	–
kMeans	–	–	–	–	7.9×	17×	23×
fuzzyCMeans	–	–	–	–	31×	49×	66×

algorithms-gpu N=256 · 4.5 sv4 (gpu) v4 (cpu)

markovClustering

72×

floydWarshall

3.4×

affinityPropagation

2.6×

1 ms10 ms100 ms1 s

All stats (table view)

group	bench	p50	avg	p99	min	samples
markovClustering	cpu	612 ms	612 ms	613 ms	611 ms	5
markovClustering	gpu	8.50 ms	9.44 ms	8.60 ms	8.30 ms	5
markovClustering	gpu first call	31.7 ms	31.7 ms	31.7 ms	31.7 ms	1
floydWarshall	cpu	23.3 ms	23.3 ms	23.3 ms	23.1 ms	5
floydWarshall	gpu	6.90 ms	6.96 ms	7.00 ms	6.80 ms	5
floydWarshall	gpu first call	20.3 ms	20.3 ms	20.3 ms	20.3 ms	1
affinityPropagation	cpu	39.8 ms	47.6 ms	40.8 ms	37.5 ms	5
affinityPropagation	gpu	15.1 ms	15.1 ms	15.2 ms	14.7 ms	5
affinityPropagation	gpu first call	25.5 ms	25.5 ms	25.5 ms	25.5 ms	1

algorithms-gpu N=512 · 33.1 sv4 (gpu) v4 (cpu)

markovClustering

300×

pageRank

1.27×

floydWarshall

11×

betweennessCentrality

3.8×

affinityPropagation

1.86×

1 ms10 ms100 ms1 s10 s

All stats (table view)

group	bench	p50	avg	p99	min	samples
markovClustering	cpu	4.78 s	4.78 s	4.79 s	4.77 s	5
markovClustering	gpu	15.9 ms	17.4 ms	16.2 ms	15.8 ms	5
markovClustering	gpu first call	25.6 ms	25.6 ms	25.6 ms	25.6 ms	1
pageRank	cpu	8.00 ms	8.14 ms	8.10 ms	7.90 ms	5
pageRank	gpu	6.30 ms	6.48 ms	6.40 ms	6.20 ms	5
pageRank	gpu first call	13.3 ms	13.3 ms	13.3 ms	13.3 ms	1
floydWarshall	cpu	179 ms	216 ms	181 ms	176 ms	5
floydWarshall	gpu	16.1 ms	16.8 ms	17.0 ms	14.7 ms	5
floydWarshall	gpu first call	18.0 ms	18.0 ms	18.0 ms	18.0 ms	1
betweennessCentrality	cpu	34.6 ms	34.9 ms	35.1 ms	34.4 ms	5
betweennessCentrality	gpu	9.10 ms	9.18 ms	9.20 ms	8.70 ms	5
betweennessCentrality	gpu first call	19.6 ms	19.6 ms	19.6 ms	19.6 ms	1
affinityPropagation	cpu	209 ms	317 ms	483 ms	204 ms	5
affinityPropagation	gpu	112 ms	112 ms	113 ms	112 ms	5
affinityPropagation	gpu first call	117 ms	117 ms	117 ms	117 ms	1

algorithms-gpu N=1,024 · 3.4 minv4 (gpu) v4 (cpu)

markovClustering

642×

pageRank

1.67×

floydWarshall

32×

betweennessCentrality

10×

affinityPropagation

5.1×

kMedoids

23×

hierarchicalClustering

2.1×

1 ms10 ms100 ms1 s10 s100 s

All stats (table view)

group	bench	p50	avg	p99	min	samples
markovClustering	cpu	30.2 s	30.2 s	30.3 s	30.1 s	5
markovClustering	gpu	47.1 ms	50.2 ms	47.3 ms	46.9 ms	5
markovClustering	gpu first call	62.0 ms	62.0 ms	62.0 ms	62.0 ms	1
pageRank	cpu	32.1 ms	42.8 ms	60.4 ms	28.8 ms	5
pageRank	gpu	19.2 ms	19.5 ms	20.1 ms	18.8 ms	5
pageRank	gpu first call	21.3 ms	21.3 ms	21.3 ms	21.3 ms	1
floydWarshall	cpu	1.35 s	1.77 s	2.57 s	1.16 s	5
floydWarshall	gpu	42.1 ms	42.9 ms	46.2 ms	39.6 ms	5
floydWarshall	gpu first call	46.5 ms	46.5 ms	46.5 ms	46.5 ms	1
betweennessCentrality	cpu	153 ms	154 ms	154 ms	148 ms	5
betweennessCentrality	gpu	15.2 ms	16.2 ms	15.5 ms	14.6 ms	5
betweennessCentrality	gpu first call	17.5 ms	17.5 ms	17.5 ms	17.5 ms	1
affinityPropagation	cpu	1.49 s	1.64 s	1.54 s	1.13 s	5
affinityPropagation	gpu	293 ms	293 ms	294 ms	290 ms	5
affinityPropagation	gpu first call	305 ms	305 ms	305 ms	305 ms	1
kMedoids	cpu	78.6 ms	82.9 ms	85.9 ms	56.9 ms	5
kMedoids	gpu	3.40 ms	3.78 ms	4.40 ms	3.20 ms	5
kMedoids	gpu first call	12.2 ms	12.2 ms	12.2 ms	12.2 ms	1
hierarchicalClustering	cpu	73.5 ms	74.3 ms	75.8 ms	72.3 ms	5
hierarchicalClustering	gpu	34.8 ms	39.7 ms	38.0 ms	33.9 ms	5
hierarchicalClustering	gpu first call	34.4 ms	34.4 ms	34.4 ms	34.4 ms	1

algorithms-gpu N=2,048 · 10.6 sv4 (gpu) v4 (cpu)

pageRank

1.69×

betweennessCentrality

18×

kMedoids

103×

hierarchicalClustering

1.89×

1 ms10 ms100 ms1 s

All stats (table view)

group	bench	p50	avg	p99	min	samples
pageRank	cpu	122 ms	122 ms	123 ms	120 ms	5
pageRank	gpu	71.9 ms	72.3 ms	73.9 ms	70.1 ms	5
pageRank	gpu first call	90.7 ms	90.7 ms	90.7 ms	90.7 ms	1
betweennessCentrality	cpu	642 ms	653 ms	657 ms	625 ms	5
betweennessCentrality	gpu	35.9 ms	38.5 ms	39.2 ms	35.2 ms	5
betweennessCentrality	gpu first call	41.6 ms	41.6 ms	41.6 ms	41.6 ms	1
kMedoids	cpu	382 ms	379 ms	407 ms	300 ms	5
kMedoids	gpu	3.70 ms	5.44 ms	4.60 ms	3.60 ms	5
kMedoids	gpu first call	5.90 ms	5.90 ms	5.90 ms	5.90 ms	1
hierarchicalClustering	cpu	307 ms	306 ms	307 ms	298 ms	5
hierarchicalClustering	gpu	162 ms	166 ms	177 ms	154 ms	5
hierarchicalClustering	gpu first call	168 ms	168 ms	168 ms	168 ms	1

algorithms-gpu N=4,096 · 23.3 sv4 (gpu) v4 (cpu)

kMeans

7.9×

fuzzyCMeans

31×

kMedoids

163×

hierarchicalClustering

1.74×

1 ms10 ms100 ms1 s10 s

All stats (table view)

group	bench	p50	avg	p99	min	samples
kMeans	cpu	31.4 ms	34.0 ms	32.3 ms	29.7 ms	5
kMeans	gpu	4.00 ms	4.24 ms	4.00 ms	3.90 ms	5
kMeans	gpu first call	12.2 ms	12.2 ms	12.2 ms	12.2 ms	1
fuzzyCMeans	cpu	157 ms	157 ms	158 ms	156 ms	5
fuzzyCMeans	gpu	5.00 ms	5.18 ms	5.50 ms	4.90 ms	5
fuzzyCMeans	gpu first call	17.8 ms	17.8 ms	17.8 ms	17.8 ms	1
kMedoids	cpu	1.74 s	1.73 s	1.78 s	1.48 s	5
kMedoids	gpu	10.7 ms	13.0 ms	11.7 ms	9.50 ms	5
kMedoids	gpu first call	12.1 ms	12.1 ms	12.1 ms	12.1 ms	1
hierarchicalClustering	cpu	1.24 s	1.25 s	1.28 s	1.21 s	5
hierarchicalClustering	gpu	714 ms	696 ms	718 ms	645 ms	5
hierarchicalClustering	gpu first call	702 ms	702 ms	702 ms	702 ms	1

algorithms-gpu N=16,384 · 4.7 sv4 (gpu) v4 (cpu)

kMeans

17×

fuzzyCMeans

49×

1 ms10 ms100 ms1 s

All stats (table view)

group	bench	p50	avg	p99	min	samples
kMeans	cpu	129 ms	129 ms	129 ms	128 ms	5
kMeans	gpu	7.70 ms	8.70 ms	8.10 ms	7.30 ms	5
kMeans	gpu first call	16.7 ms	16.7 ms	16.7 ms	16.7 ms	1
fuzzyCMeans	cpu	613 ms	619 ms	615 ms	610 ms	5
fuzzyCMeans	gpu	12.4 ms	12.5 ms	12.4 ms	11.7 ms	5
fuzzyCMeans	gpu first call	21.1 ms	21.1 ms	21.1 ms	21.1 ms	1

algorithms-gpu N=65,536 · 19.9 sv4 (gpu) v4 (cpu)

kMeans

23×

fuzzyCMeans

66×

10 ms100 ms1 s10 s

All stats (table view)

group	bench	p50	avg	p99	min	samples
kMeans	cpu	628 ms	636 ms	632 ms	617 ms	5
kMeans	gpu	27.4 ms	27.1 ms	27.4 ms	26.0 ms	5
kMeans	gpu first call	68.5 ms	68.5 ms	68.5 ms	68.5 ms	1
fuzzyCMeans	cpu	2.54 s	2.55 s	2.56 s	2.51 s	5
fuzzyCMeans	gpu	38.5 ms	39.4 ms	38.8 ms	37.3 ms	5
fuzzyCMeans	gpu first call	82.5 ms	82.5 ms	82.5 ms	82.5 ms	1