Cytoscape.js GPU benchmarks

v4 executor 'cpu' vs 'gpu' · times are per-call p50 · 2026-08-12 23:08 · c2d24c2e (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	c2d24c2e docs: the summary goes day by day, from the commit record
Machine id	5cf3f79c

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

markovClustering @ 1.024k

646×

markovClustering @ 512

291×

kMedoids @ 4.096k

142×

markovClustering @ 256

69×

fuzzyCMeans @ 65.536k

59×

kMedoids @ 2.048k

59×

fuzzyCMeans @ 16.384k

49×

fuzzyCMeans @ 4.096k

32×

floydWarshall @ 1.024k

30×

kMeans @ 65.536k

25×

kMedoids @ 1.024k

23×

betweennessCentrality @ 2.048k

18×

kMeans @ 16.384k

17×

floydWarshall @ 512

12×

betweennessCentrality @ 1.024k

9.7×

kMeans @ 4.096k

7.8×

affinityPropagation @ 1.024k

3.9×

betweennessCentrality @ 512

3.9×

floydWarshall @ 256

3.4×

affinityPropagation @ 256

2.4×

affinityPropagation @ 512

1.87×

hierarchicalClustering @ 1.024k

1.04×

hierarchicalClustering @ 4.096k

0.96×

hierarchicalClustering @ 2.048k

0.91×

pageRankDense @ 1.024k

0.19×

pageRankDense @ 2.048k

0.19×

pageRankDense @ 512

0.14×

pageRank @ 512

0.10×

pageRank @ 1.024k

0.02×

pageRank @ 2.048k

0.01×

0.001×0.01×0.1×1×10×100×1000×

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	69×	291×	646×	–	–	–	–
floydWarshall	3.4×	12×	30×	–	–	–	–
affinityPropagation	2.4×	1.87×	3.9×	–	–	–	–
pageRank	–	0.10×	0.02×	0.01×	–	–	–
pageRankDense	–	0.14×	0.19×	0.19×	–	–	–
betweennessCentrality	–	3.9×	9.7×	18×	–	–	–
kMedoids	–	–	23×	59×	142×	–	–
hierarchicalClustering	–	–	1.04×	0.91×	0.96×	–	–
kMeans	–	–	–	–	7.8×	17×	25×
fuzzyCMeans	–	–	–	–	32×	49×	59×

algorithms-gpu N=256 · 4.6 sv4 (gpu) v4 (cpu)

markovClustering

69×

floydWarshall

3.4×

affinityPropagation

2.4×

1 ms10 ms100 ms1 s

All stats (table view)

group	bench	p50	avg	p99	min	samples
markovClustering	cpu	625 ms	627 ms	627 ms	622 ms	5
markovClustering	gpu	9.00 ms	9.18 ms	9.00 ms	8.50 ms	5
markovClustering	gpu first call	27.4 ms	27.4 ms	27.4 ms	27.4 ms	1
floydWarshall	cpu	23.3 ms	23.3 ms	23.4 ms	23.3 ms	5
floydWarshall	gpu	6.90 ms	6.82 ms	7.00 ms	6.50 ms	5
floydWarshall	gpu first call	29.6 ms	29.6 ms	29.6 ms	29.6 ms	1
affinityPropagation	cpu	38.4 ms	40.5 ms	38.8 ms	38.3 ms	5
affinityPropagation	gpu	15.8 ms	16.0 ms	15.9 ms	15.3 ms	5
affinityPropagation	gpu first call	26.0 ms	26.0 ms	26.0 ms	26.0 ms	1

algorithms-gpu N=512 · 33.5 sv4 (gpu) v4 (cpu)

markovClustering

291×

pageRank

0.10×

pageRankDense

0.14×

floydWarshall

12×

betweennessCentrality

3.9×

affinityPropagation

1.87×

100 µs1 ms10 ms100 ms1 s10 s

All stats (table view)

group	bench	p50	avg	p99	min	samples
markovClustering	cpu	4.89 s	4.89 s	4.90 s	4.88 s	5
markovClustering	gpu	16.8 ms	17.9 ms	17.0 ms	15.9 ms	5
markovClustering	gpu first call	27.2 ms	27.2 ms	27.2 ms	27.2 ms	1
pageRank	cpu	600 µs	560 µs	700 µs	200 µs	5
pageRank	gpu	6.30 ms	6.54 ms	7.00 ms	6.10 ms	5
pageRank	gpu first call	13.6 ms	13.6 ms	13.6 ms	13.6 ms	1
pageRankDense	cpu	1.00 ms	980 µs	1.00 ms	900 µs	5
pageRankDense	gpu	6.90 ms	6.82 ms	6.90 ms	6.60 ms	5
pageRankDense	gpu first call	30.0 ms	30.0 ms	30.0 ms	30.0 ms	1
floydWarshall	cpu	180 ms	210 ms	188 ms	178 ms	5
floydWarshall	gpu	15.1 ms	15.6 ms	15.4 ms	14.9 ms	5
floydWarshall	gpu first call	17.5 ms	17.5 ms	17.5 ms	17.5 ms	1
betweennessCentrality	cpu	34.4 ms	34.5 ms	34.5 ms	34.1 ms	5
betweennessCentrality	gpu	8.80 ms	9.18 ms	9.00 ms	8.60 ms	5
betweennessCentrality	gpu first call	20.0 ms	20.0 ms	20.0 ms	20.0 ms	1
affinityPropagation	cpu	209 ms	266 ms	212 ms	201 ms	5
affinityPropagation	gpu	112 ms	112 ms	113 ms	111 ms	5
affinityPropagation	gpu first call	115 ms	115 ms	115 ms	115 ms	1

algorithms-gpu N=1,024 · 3.5 minv4 (gpu) v4 (cpu)

markovClustering

646×

pageRank

0.02×

pageRankDense

0.19×

floydWarshall

30×

betweennessCentrality

9.7×

affinityPropagation

3.9×

kMedoids

23×

hierarchicalClustering

1.04×

100 µs1 ms10 ms100 ms1 s10 s100 s

All stats (table view)

group	bench	p50	avg	p99	min	samples
markovClustering	cpu	30.3 s	30.3 s	30.4 s	30.3 s	5
markovClustering	gpu	46.9 ms	50.4 ms	54.1 ms	46.1 ms	5
markovClustering	gpu first call	62.2 ms	62.2 ms	62.2 ms	62.2 ms	1
pageRank	cpu	300 µs	460 µs	700 µs	300 µs	5
pageRank	gpu	18.0 ms	18.2 ms	18.0 ms	17.4 ms	5
pageRank	gpu first call	19.2 ms	19.2 ms	19.2 ms	19.2 ms	1
pageRankDense	cpu	3.90 ms	3.84 ms	3.90 ms	3.70 ms	5
pageRankDense	gpu	20.7 ms	20.7 ms	20.8 ms	20.2 ms	5
pageRankDense	gpu first call	107 ms	107 ms	107 ms	107 ms	1
floydWarshall	cpu	1.41 s	1.79 s	2.57 s	1.18 s	5
floydWarshall	gpu	46.6 ms	51.1 ms	48.1 ms	38.1 ms	5
floydWarshall	gpu first call	51.1 ms	51.1 ms	51.1 ms	51.1 ms	1
betweennessCentrality	cpu	149 ms	153 ms	151 ms	148 ms	5
betweennessCentrality	gpu	15.4 ms	16.5 ms	15.4 ms	15.2 ms	5
betweennessCentrality	gpu first call	17.5 ms	17.5 ms	17.5 ms	17.5 ms	1
affinityPropagation	cpu	1.15 s	1.58 s	2.30 s	1.05 s	5
affinityPropagation	gpu	295 ms	295 ms	296 ms	290 ms	5
affinityPropagation	gpu first call	292 ms	292 ms	292 ms	292 ms	1
kMedoids	cpu	72.6 ms	81.0 ms	93.1 ms	51.4 ms	5
kMedoids	gpu	3.20 ms	3.58 ms	4.10 ms	3.10 ms	5
kMedoids	gpu first call	13.5 ms	13.5 ms	13.5 ms	13.5 ms	1
hierarchicalClustering	cpu	28.0 ms	29.3 ms	35.0 ms	23.8 ms	5
hierarchicalClustering	gpu	27.0 ms	27.0 ms	27.1 ms	26.8 ms	5
hierarchicalClustering	gpu first call	36.0 ms	36.0 ms	36.0 ms	36.0 ms	1

algorithms-gpu N=2,048 · 9.1 sv4 (gpu) v4 (cpu)

pageRank

0.01×

pageRankDense

0.19×

betweennessCentrality

18×

kMedoids

59×

hierarchicalClustering

0.91×

100 µs1 ms10 ms100 ms1 s

All stats (table view)

group	bench	p50	avg	p99	min	samples
pageRank	cpu	600 µs	620 µs	600 µs	500 µs	5
pageRank	gpu	72.5 ms	72.0 ms	72.8 ms	68.2 ms	5
pageRank	gpu first call	85.3 ms	85.3 ms	85.3 ms	85.3 ms	1
pageRankDense	cpu	15.4 ms	15.4 ms	15.5 ms	15.2 ms	5
pageRankDense	gpu	82.5 ms	82.4 ms	83.0 ms	81.1 ms	5
pageRankDense	gpu first call	380 ms	380 ms	380 ms	380 ms	1
betweennessCentrality	cpu	640 ms	661 ms	715 ms	611 ms	5
betweennessCentrality	gpu	36.5 ms	38.5 ms	38.0 ms	35.6 ms	5
betweennessCentrality	gpu first call	42.7 ms	42.7 ms	42.7 ms	42.7 ms	1
kMedoids	cpu	271 ms	309 ms	382 ms	228 ms	5
kMedoids	gpu	4.60 ms	5.82 ms	4.70 ms	3.60 ms	5
kMedoids	gpu first call	5.90 ms	5.90 ms	5.90 ms	5.90 ms	1
hierarchicalClustering	cpu	98.6 ms	99.0 ms	99.2 ms	98.0 ms	5
hierarchicalClustering	gpu	108 ms	108 ms	108 ms	105 ms	5
hierarchicalClustering	gpu first call	106 ms	106 ms	106 ms	106 ms	1

algorithms-gpu N=4,096 · 17.1 sv4 (gpu) v4 (cpu)

kMeans

7.8×

fuzzyCMeans

32×

kMedoids

142×

hierarchicalClustering

0.96×

1 ms10 ms100 ms1 s10 s

All stats (table view)

group	bench	p50	avg	p99	min	samples
kMeans	cpu	30.6 ms	30.9 ms	30.8 ms	29.7 ms	5
kMeans	gpu	3.90 ms	3.88 ms	4.00 ms	3.70 ms	5
kMeans	gpu first call	12.2 ms	12.2 ms	12.2 ms	12.2 ms	1
fuzzyCMeans	cpu	152 ms	152 ms	152 ms	151 ms	5
fuzzyCMeans	gpu	4.80 ms	5.00 ms	5.40 ms	4.60 ms	5
fuzzyCMeans	gpu first call	18.0 ms	18.0 ms	18.0 ms	18.0 ms	1
kMedoids	cpu	1.71 s	1.76 s	1.87 s	1.63 s	5
kMedoids	gpu	12.0 ms	13.1 ms	12.2 ms	9.80 ms	5
kMedoids	gpu first call	12.0 ms	12.0 ms	12.0 ms	12.0 ms	1
hierarchicalClustering	cpu	422 ms	424 ms	425 ms	415 ms	5
hierarchicalClustering	gpu	439 ms	448 ms	459 ms	436 ms	5
hierarchicalClustering	gpu first call	493 ms	493 ms	493 ms	493 ms	1

algorithms-gpu N=16,384 · 4.6 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	124 ms	126 ms	129 ms	122 ms	5
kMeans	gpu	7.30 ms	7.30 ms	7.30 ms	7.20 ms	5
kMeans	gpu first call	17.4 ms	17.4 ms	17.4 ms	17.4 ms	1
fuzzyCMeans	cpu	600 ms	604 ms	607 ms	600 ms	5
fuzzyCMeans	gpu	12.3 ms	16.8 ms	12.3 ms	11.2 ms	5
fuzzyCMeans	gpu first call	22.0 ms	22.0 ms	22.0 ms	22.0 ms	1

algorithms-gpu N=65,536 · 19.6 sv4 (gpu) v4 (cpu)

kMeans

25×

fuzzyCMeans

59×

10 ms100 ms1 s10 s

All stats (table view)

group	bench	p50	avg	p99	min	samples
kMeans	cpu	611 ms	600 ms	616 ms	570 ms	5
kMeans	gpu	24.3 ms	27.4 ms	24.7 ms	24.2 ms	5
kMeans	gpu first call	70.3 ms	70.3 ms	70.3 ms	70.3 ms	1
fuzzyCMeans	cpu	2.53 s	2.56 s	2.56 s	2.52 s	5
fuzzyCMeans	gpu	42.6 ms	41.2 ms	42.8 ms	37.4 ms	5
fuzzyCMeans	gpu first call	84.3 ms	84.3 ms	84.3 ms	84.3 ms	1