Clusters appear across the universe and across disciplines: galaxy clusters in space, dark-matter tests like the Bullet Cluster, and clustering for data in medicine and biology. As you move through the page, you’ll see examples ranging from tumor resistance and hippocampal engrams to biofilms and protein droplets. You’ll also compare real cluster computing with classic methods such as k-means, DBSCAN, hierarchical clustering, and Gaussian mixture models.
Key Takeaways
- The Virgo Cluster, the heart of the Local Supercluster, spans 15 million light-years and contains 1,300 to 2,000 galaxies.
- Abell 1689, a massive galaxy cluster, acts as a gravitational lens magnifying distant galaxies by a factor of 30.
- The Coma Cluster hosts over 1,000 galaxies within a diameter of 6 million light-years.
- Cancer stem cell clusters in tumors show 10-fold higher resistance to chemo.
- Neuronal clusters in hippocampus form 1,000-cell engrams for memory.
- Bacterial biofilms form clusters with 10^9 cells/cm² density.
- IBM's Summit supercomputer cluster achieves 200 petaflops peak performance with 4,608 nodes.
- Fugaku cluster in Japan reaches 442 petaflops on HPL benchmark with 158,976 nodes.
- Frontier exascale cluster at ORNL delivers 1.102 exaflops with AMD GPUs across 9,472 nodes.
- Silicon Valley hosts 40% of US venture capital-funded tech firms in cluster.
- Route 128 cluster in Boston generated 15% of US software revenue in 1990s.
- Hollywood film cluster employs 250,000 workers, producing $50B annually.
- K-means clustering on Iris dataset typically yields 3 optimal clusters with silhouette score 0.55.
- DBSCAN on synthetic moons dataset finds 2 clusters with eps=0.3, min_samples=10.
- Hierarchical clustering dendrogram for 150 samples cuts at 3 clusters with cophenetic correlation 0.93.
From galaxy clusters to algorithms and supercomputers, clustering reveals hidden structure everywhere.
Related reading
01 · Category
Astronomical30 stats
01
The Virgo Cluster, the heart of the Local Supercluster, spans 15 million light-years and contains 1,300 to 2,000 galaxies.
02
Abell 1689, a massive galaxy cluster, acts as a gravitational lens magnifying distant galaxies by a factor of 30.
03
The Coma Cluster hosts over 1,000 galaxies within a diameter of 6 million light-years.
04
Bullet Cluster provides evidence for dark matter with a mass separation of 5 arcminutes between baryonic and gravitational mass.
05
El Gordo cluster has a total mass of 3 quadrillion solar masses at redshift z=0.87.
06
The Perseus Cluster's central black hole outputs 10^44 ergs per second in radio luminosity.
07
MACS J0416.1-2403 cluster lenses light from 9 background galaxies with magnification up to 30 times.
08
The Hercules–Corona Borealis Great Wall spans 10 billion light-years, containing multiple clusters.
09
Abell 2744, Pandora's Cluster, has a complex mass distribution from four colliding subclusters.
10
The Shapley Supercluster contains over 20 galaxy clusters within 650 million light-years.
11
IC 1101 in Abell 2020 is the largest known galaxy in a cluster, spanning 6 million light-years.
12
The Centaurus Cluster has a velocity dispersion of 1,200 km/s for its galaxies.
13
Zwicky 3146 cluster shows a dark matter core misalignment of 19 degrees.
14
The Phoenix Cluster has a central galaxy star formation rate of 740 solar masses per year.
15
Abell 2218 gravitational lens arcs span 40 arcseconds from lensing.
16
The Local Group contains over 54 galaxies, including cluster-like dwarf associations.
17
Stephan's Quintet is a compact group with 5 galaxies interacting over 100 million light-years.
18
The Rich Cluster CIZA J2242.8+5301 has a shock front moving at 1,600 km/s.
19
Fornax Cluster has 58 confirmed member galaxies within 2 degrees.
20
The Eridanus Supervoid spans 1 billion light-years, affecting cluster distributions.
21
Abell 520 shows a dark matter "bullet" collision remnant.
22
The Musca Cluster in the Milky Way has 700 young stars.
23
NGC 6334 cluster forms 100 stars per year in its molecular cloud.
24
Westerlund 1 super star cluster has 3,000 red supergiants.
25
Arches Cluster near Galactic Center has 10,000 stars within 0.5 pc.
26
Quintuplet Cluster has a core density of 75,000 stars per pc³.
27
Cygnus OB2 association has over 1,000 OB stars.
28
RCW 38 embedded cluster has 500 young stars.
29
Trumpler 14 cluster in Carina Nebula has 100 O/B stars.
30
Lagoon Nebula cluster (NGC 6530) has 1,000 members.
Interpretation
Astronomical Interpretation
Across astronomical clusters, the sheer scale of their mass and structure is striking, from the Virgo Cluster’s 1,300 to 2,000 galaxies over 15 million light-years to the Bullet Cluster’s 5 arcminutes separation between baryonic and gravitational mass, showing how gravity reveals the universe’s hidden components on vast scales.
More related reading
02 · Category
Biology20 stats
01
Cancer stem cell clusters in tumors show 10-fold higher resistance to chemo.
02
Neuronal clusters in hippocampus form 1,000-cell engrams for memory.
03
Bacterial biofilms form clusters with 10^9 cells/cm² density.
04
Protein clusters in phase-separated droplets number 100s per cell.
05
Lymphocyte clusters in lymph nodes peak at 500 cells per follicle.
06
Mitochondrial clusters in muscle fibers span 50 μm lengths.
07
Actin clusters drive filopodia with 200 monomers per tip.
08
Synaptic vesicle clusters contain 100-200 vesicles per active zone.
09
Chromatin clusters in nucleus form topologically associating domains (TADs) averaging 1 Mb.
10
Endoplasmic reticulum clusters export 1,000 proteins/sec per cell.
11
Golgi stacks cluster into 50-100 cisternae per apparatus.
12
Peroxisome clusters in yeast number 20-50 per cell under stress.
13
RNA granules cluster mRNAs into P-bodies with 50-100 transcripts each.
14
Tight junction protein clusters span 1 μm with 10^4 claudins.
15
Desmosome clusters in epithelia anchor 1,000 cadherins per plaque.
16
Gap junction clusters conduct 100 pS per connexon.
17
Spindle pole body clusters microtubules into 30-40 per pole in yeast.
18
Centrosome clusters nucleate 100 microtubules in interphase.
19
Immune synapse clusters TCRs into 100 nm domains.
20
Caveolae clusters invaginate plasma membrane with 100-150 proteins each.
Interpretation
Biology Interpretation
Across biology, these cluster statistics show a striking scale pattern where biological systems range from 10-fold chemo-resistant cancer stem cell clusters to neuron engrams that can reach 1,000 cells and bacterial biofilm clusters packed at 10^9 cells per cm², highlighting how clustered organization shapes function from cellular survival to memory and tissue-level structure.
More related reading
03 · Category
Computing28 stats
01
IBM's Summit supercomputer cluster achieves 200 petaflops peak performance with 4,608 nodes.
02
Fugaku cluster in Japan reaches 442 petaflops on HPL benchmark with 158,976 nodes.
03
Frontier exascale cluster at ORNL delivers 1.102 exaflops with AMD GPUs across 9,472 nodes.
04
LUMI supercomputer cluster in Finland has 2,752 nodes with 576 GPUs each.
05
MareNostrum 4 cluster peaks at 11.0 petaflops with 3,674 compute nodes.
06
AWS Graviton3-based cluster instances provide up to 40% better price-performance.
07
Kubernetes clusters manage 1.5 billion container workloads daily across enterprises.
08
Hadoop clusters process 100 petabytes of data daily in large deployments.
09
Google Borealis cluster runs 1 million jobs per day on AI workloads.
10
OpenMPI clusters scale to 100,000 cores with <1% overhead.
11
SLURM workload manager schedules 90% of TOP500 clusters.
12
NVLink in GPU clusters provides 600 GB/s bidirectional bandwidth per GPU.
13
InfiniBand HDR clusters achieve 200 Gb/s per port latency under 1 μs.
14
Azure Batch clusters support up to 200,000 nodes for HPC.
15
AWS ParallelCluster deploys clusters with 100,000 cores in under 10 minutes.
16
Google Cloud HPC Toolkit builds clusters with 95% utilization.
17
CERN's LHC computing grid clusters process 300 million events per second.
18
Tianhe-2A cluster sustained 33.86 petaflops with Intel Xeon Phi.
19
Piz Daint cluster achieved 21.2 petaflops with 5,272 nodes.
20
Blue Waters cluster ran 10 million core-hours monthly.
21
K computer cluster used 88,128 nodes for 10.51 petaflops.
22
Sunway TaihuLight cluster hit 93 petaflops with SW26010 processors.
23
Sierra cluster at LLNL delivers 94.6 petaflops with 4,320 nodes.
24
148.6 exaflops peak HPL performance for Summit in 2018
25
442.0 exaflops peak HPL performance for Fugaku in 2020
26
591.0 exaflops peak HPL performance for Frontier in 2023
27
2.0 exaflops peak HPL performance for Summit in 2019
28
0.0 exaflops peak HPL performance for Frontier in 2022
Interpretation
Computing Interpretation
Across computing clusters, the push for scale and speed is dramatic, with peak performance rising from 200 petaflops on IBM’s Summit to 442 petaflops on Japan’s Fugaku and then jumping to 1.102 exaflops on ORNL’s Frontier, showing how node and accelerator growth drives the next wave of computing power.
More related reading
report visual · Comparison
Peak HPL Performance Rise in Major Supercomputer Clusters
Across Summit, Fugaku, and Frontier, peak HPL performance trends upward, with Frontier leading by 2023 and a clear step-up from earlier leaders as rankings shift over time.
591.0 exaflops peak HPL performance for Frontier in 2023591.0 exaflops
442.0 exaflops peak HPL performance for Fugaku in 2020
442.0 exaflops
148.6 exaflops peak HPL performance for Summit in 2018
148.6 exaflops
source-verifiedolcf.ornl.gov · r-ccs.riken.jp2023
04 · Category
Economics21 stats
01
Silicon Valley hosts 40% of US venture capital-funded tech firms in cluster.
02
Route 128 cluster in Boston generated 15% of US software revenue in 1990s.
03
Hollywood film cluster employs 250,000 workers, producing $50B annually.
04
Detroit auto cluster peaked at 25% global car production in 1950s.
05
Emilia-Romagna Italy's ceramic district exports €5B yearly from 300 firms.
06
Third Italy textile clusters contribute 12% to national exports.
07
Bangalore IT cluster employs 1.5M workers, 40% of India's software exports.
08
Shenzhen electronics cluster produces 90% of world's consumer drones.
09
Guadalajara Mexico's maquiladora cluster assembles 50% of US TVs.
10
Singapore biotech cluster invests $15B in R&D hubs.
11
Wall Street finance cluster handles 40% of global trades.
12
Nashville music cluster generates $15B economic impact annually.
13
Austin tech cluster grew 50% in startups from 2015-2020.
14
Pearl River Delta manufacturing cluster outputs 10% of world goods.
15
Basque Country machine tool cluster leads Europe with 30% market share.
16
Montreal aerospace cluster exports $15B, employs 40,000.
17
Tuscany leather cluster in Italy produces 35M pairs of shoes yearly.
18
Salt Lake City medical devices cluster has 500 firms, $10B output.
19
Clusters boost firm productivity by 15-20% via knowledge spillovers.
20
US biotech clusters employ 800,000, R&D spend $100B/year.
21
Porter's diamond model shows clusters increase competitiveness by 33%.
Interpretation
Economics Interpretation
Across these economics-focused clusters, deep industry concentrations are driving outsized national and global impact, such as Silicon Valley accounting for 40% of US venture capital-funded tech firms and Route 128 delivering 15% of US software revenue in the 1990s.
More related reading
05 · Category
Statistics20 stats
01
K-means clustering on Iris dataset typically yields 3 optimal clusters with silhouette score 0.55.
02
DBSCAN on synthetic moons dataset finds 2 clusters with eps=0.3, min_samples=10.
03
Hierarchical clustering dendrogram for 150 samples cuts at 3 clusters with cophenetic correlation 0.93.
04
Gaussian Mixture Model on blobs data fits 4 components with BIC=-7000.
05
Spectral clustering on Swiss roll dataset recovers 3 manifolds with Laplacian eigenmaps.
06
Agglomerative clustering with Ward linkage on random data gives 5 clusters, inertia=150.
07
Birch clustering on 100,000 points scales to threshold=0.5, branching=50 clusters.
08
OPTICS clustering extracts 4 clusters from noisy data with xi=0.05.
09
Mean Shift clustering on 2 moons finds 2 modes with bandwidth=0.65.
10
Affinity Propagation on 300 samples prefers 12 exemplars.
11
In customer segmentation, RFM clustering yields 5 segments covering 92% variance.
12
Image segmentation with SLIC superpixels generates 400 clusters per 512x512 image.
13
Topic modeling LDA on 20NG extracts 20 clusters with coherence 0.45.
14
Anomaly detection isolation forest flags 10% outliers in 100k samples.
15
t-SNE visualization clusters MNIST digits into 10 manifolds with perplexity=30.
16
UMAP reduces 1M GloVe embeddings to 2D with 15 clusters, n_neighbors=15.
17
HDBSCAN on spatial data finds 50 clusters with min_cluster_size=10.
18
Fuzzy C-Means on medical images partitions into 4 clusters with membership 0.8.
19
Self-Organizing Maps (SOM) on stocks data forms 10x10 grid clusters.
20
Growing Neural Gas clusters streaming data into 200 nodes online.
Interpretation
Statistics Interpretation
Across these Statistics examples, multiple clustering methods consistently settle on small, interpretable group counts, such as K means giving 3 clusters, hierarchical cutting into 3 clusters with a high cophenetic correlation of 0.93, and spectral clustering recovering 3 manifolds, suggesting that even varied algorithms often converge on similarly structured solutions.
Reference
Cite This Report
This report is designed to be cited. We maintain stable URLs and versioned verification dates. Copy the format appropriate for your publication below.
APA
Isabelle Moreau. (2026, February 13). Clusters Statistics. Gitnux. https://gitnux.org/clusters-statistics
MLA
Isabelle Moreau. "Clusters Statistics." Gitnux, 13 Feb 2026, https://gitnux.org/clusters-statistics.
Chicago
Isabelle Moreau. 2026. "Clusters Statistics." Gitnux. https://gitnux.org/clusters-statistics.
Sources & references
5 datasets cited across this report · attribution is report-level
+2 additional datasets cited (not shown individually)