Skip to main content

Hetzner CPX21

Hillsboro, OR

Benchmark complete
Finished Oct 17, 2026
Geekbench 7 single-core
1,390
Geekbench 7 multi-core
Skipped
Disk 4K IOPS
56,416
fio random read and write, 50/50
Monthly price at test time
$37.49
Same price today

Server

  • Provider
    Hetzner
  • Plan
    CPX21
  • Location
    Hillsboro, OR
  • vCPU
    3
  • RAM
    4 GB
  • Storage
    80 GB NVMe
  • Network
    Unknown
  • OS image
    ubuntu-24.04
  • Trigger
    Scheduled
  • Queued at
    Oct 17, 2026, 06:35:00 UTC
Visit Hetzner

Latest runs of this plan

Full history of this plan

CPU

Version Single-core Multi-core
Geekbench 7 1,390 Skipped Geekbench Browser

Multi-core was skipped because Geekbench 7 needs about 3.8 GB of free memory to run 3 threads, and this server didn't have that much. Single-core needs much less, so it still ran.

Scores from different Geekbench versions aren't comparable with each other. Charts only use Geekbench 7.

Geekbench 7 workloads

Workload Single-core Multi-core
File Compression
1,428
76th of 216
Single-core only
Navigation
1,453
68th of 216
Single-core only
HTML5 Browser
1,337
74th of 216
Single-core only
PDF Viewer
1,456
78th of 216
Single-core only
Photo Library
1,301
71st of 216
Single-core only
Clang
1,140
102nd of 216
Single-core only
Text Processing
1,207
87th of 216
Single-core only
Asset Compression
1,382
86th of 216
Single-core only
HDR
1,365
73rd of 216
Single-core only
Photo Editor
1,653
60th of 216
Single-core only
Ray Tracer
1,321
78th of 216
Single-core only
Structure from Motion
1,721
45th of 216
Single-core only
Game Physics
1,303
79th of 216
Single-core only
Video Encoder
1,320
71st of 216
Single-core only
Audio Encoder
1,439
69th of 216
Single-core only
Video Player
1,538
64th of 216
Single-core only

Rank among the 216 runs with Geekbench 7 workload scores. Bars are relative to the best run. Geekbench doesn't run every workload multi-core, so some rows only have a single-core score.

Memory, TLS and scheduler

Memory bandwidth (STREAM)

Test Result Runs Spread
Copy higher is better 63.2 GB/s 3 20.3%
Scale higher is better 49.3 GB/s 3 19.7%
Add higher is better 52.5 GB/s 3 20.3%
Triad higher is better 53.7 GB/s 3 21.4%

Average of 3 runs. Spread is the standard deviation as a share of the average.

TLS and crypto (OpenSSL)

Test Result Runs Spread
RSA 4096 signs higher is better 639/s 3 0.452%
RSA 4096 verifies higher is better 42,301/s 3 0.83%
SHA-256 higher is better 2.4 GB/s 3 0.243%
AES-256-GCM higher is better 11.3 GB/s 3 0.437%
ChaCha20-Poly1305 higher is better 6.4 GB/s 3 0.427%

Average of 3 runs. Spread is the standard deviation as a share of the average.

Scheduler latency (schbench)

Test Result Runs Spread
Request latency, median lower is better 4.69 ms 3 0.197%
Requests per second, 90th percentile higher is better 623/s 3 0.321%
Request latency, 99.9th percentile lower is better 9.37 ms 3 11.5%
Requests per second, average higher is better 616/s 3 0.152%

Average of 3 runs. Spread is the standard deviation as a share of the average.

Disk

Block size Read Write Total
4K 110 MB/s (28,171 IOPS) 110.3 MB/s (28,245 IOPS) 220.4 MB/s (56,416 IOPS)
64K 1010.2 MB/s (16,163 IOPS) 1015.5 MB/s (16,248 IOPS) 1.98 GB/s (32,411 IOPS)
512K 1.67 GB/s (3,419 IOPS) 1.76 GB/s (3,600 IOPS) 3.43 GB/s (7,019 IOPS)
1M 1.86 GB/s (1,907 IOPS) 1.99 GB/s (2,034 IOPS) 3.85 GB/s (3,941 IOPS)

fio random read and write, 50/50, run by YABS against the boot disk.

Network

IPv4 transfer speed

Loading...
Target Protocol Send Receive
Clouvider, London, UK (10G) IPv4 1.17 Gbps 1.47 Gbps
Eranium, Amsterdam, NL (100G) IPv4 1.23 Gbps 1.41 Gbps
Leaseweb, Singapore, SG (10G) IPv4 1.01 Gbps 1.39 Gbps
Clouvider, Los Angeles, CA, US (10G) IPv4 4.41 Gbps 7.92 Gbps
Leaseweb, NYC, NY, US (10G) IPv4 2.41 Gbps 3.11 Gbps

Public iperf3 servers vary in capacity and availability. Compare a target only with itself over time.

Datastores

PostgreSQL

fsync speed (pg_test_fsync)

Method Operations per second Time per operation
One 8 kB write
open_datasync 3,438 0.291 ms
fdatasync 3,291 0.304 ms
fsync 1,599 0.625 ms
fsync_writethrough Not supported here n/a
open_sync 2,114 0.473 ms
Two 8 kB writes
open_datasync 1,780 0.562 ms
fdatasync 3,242 0.308 ms
fsync 1,943 0.515 ms
fsync_writethrough Not supported here n/a
open_sync 1,004 0.996 ms
open_sync at different write sizes
1 * 16kB open_sync write 1,971 0.507 ms
2 * 8kB open_sync writes 1,022 0.978 ms
4 * 4kB open_sync writes 510 1.96 ms
8 * 2kB open_sync writes 217 4.61 ms
16 * 1kB open_sync writes 121 8.25 ms
fsync on a non-write descriptor
write, fsync, close 1,870 0.535 ms
write, close, fsync 1,520 0.658 ms
Non-synced 8 kB writes
write 1,318,692 0.001 ms

pg_test_fsync against the Docker volume Postgres keeps its data on. fdatasync is Linux's default wal_sync_method.

pgbench at scale 10

Workload Clients Transactions/s Median p95 p99 p99.9 Failed
Read/write (TPC-B like) 1 999 0.864 ms 1.59 ms 1.9 ms 4.39 ms 0
Read/write (TPC-B like) 8 5,279 1.36 ms 2.67 ms 3.91 ms 6.34 ms 0
Read/write (TPC-B like) 32 5,798 4.22 ms 13.5 ms 22.9 ms 38.1 ms 0
Read only 1 16,039 0.056 ms 0.093 ms 0.125 ms 0.19 ms 0
Read only 8 77,475 0.094 ms 0.172 ms 0.22 ms 0.363 ms 0
Read only 32 94,315 0.294 ms 0.499 ms 0.655 ms 1.13 ms 0

Scale 10 (1,000,000 accounts), loaded in 1.13 s. 60 s per run, prepared statements. Latency comes from sampled transaction logs, and p99.9 is left out when fewer than 1,000 were sampled.

pgbench at scale 100

Workload Clients Transactions/s Median p95 p99 p99.9 Failed
Read/write (TPC-B like) 1 869 1.04 ms 1.83 ms 2.74 ms 4.52 ms 0
Read/write (TPC-B like) 8 5,227 1.37 ms 2.61 ms 4.84 ms 9.32 ms 0
Read/write (TPC-B like) 32 7,840 3.73 ms 6.77 ms 9.58 ms 18.3 ms 0
Read only 1 12,015 0.075 ms 0.13 ms 0.186 ms 0.43 ms 0
Read only 8 60,788 0.121 ms 0.218 ms 0.281 ms 0.615 ms 0
Read only 32 71,629 0.4 ms 0.689 ms 0.921 ms 1.66 ms 0

Scale 100 (10,000,000 accounts), loaded in 11.1 s. 60 s per run, prepared statements. Latency comes from sampled transaction logs, and p99.9 is left out when fewer than 1,000 were sampled.

MySQL

sysbench OLTP

Workload Threads Transactions/s Queries/s Median p95 p99 Errors
Read/write 1 205 4,110 4.41 ms 7.43 ms 11.9 ms 0
Read/write 8 1,036 20,716 7.43 ms 11.7 ms 16.1 ms 0
Read/write 32 1,316 26,324 23.9 ms 38.2 ms 47.5 ms 0
Read only 1 641 10,261 1.42 ms 2.48 ms 3.13 ms 0
Read only 8 2,749 43,980 2.76 ms 4.65 ms 5.47 ms 0
Read only 32 2,772 44,356 11 ms 16.1 ms 18.6 ms 0

10 tables of 100,000 rows, loaded in 5.6 s. 60 s per run, one connection per thread.

Redis

memtier

Mix Operations/s Hits/s Median p99 p99.9
1:10 set:get 105,532 95,938 0.575 ms 1.2 ms 2.02 ms
1:1 set:get 100,891 50,445 0.599 ms 1.24 ms 2.1 ms

1,000,000 keys of 32 bytes. 64 connections over 2 threads, 60 s per ratio.

Web frameworks

Framework Fortunes req/s Single query req/s JSON req/s Fortunes p99 at 256 connections
Next.js nextjs 1,014 3,556 5,092 400 ms
Express express-postgresjs 11,568 16,758 51,896 40.3 ms
Fastify fastify-postgres 11,572 17,144 57,200 39.8 ms
NestJS nestjs 7,026 7,909 26,995 63.2 ms
Nuxt nuxt 10,325 15,639 53,380 59.5 ms
FastAPI fastapi 5,422 6,731 27,667 227 ms
Django django-postgresql 2,643 3,038 7,321 145 ms
Rails rails 4,551 6,028 24,051 99.9 ms
Laravel laravel 1,174 1,241 2,344 238 ms
Symfony symfony 1,570 1,724 4,548 177 ms

Peak requests per second across the 64 and 256 connection levels, measured with wrk. Each test has to pass the TechEmpower response checks first, or it's marked invalid.

Web apps

WooCommerce

Page mix

Step Requests/s Median p95 p99 Error rate
8 virtual users
Home page 3.5 460 ms 550 ms 582 ms 0%
Shop page 3.5 607 ms 737 ms 769 ms 0%
Category page 3.5 386 ms 641 ms 721 ms 0%
Product page 3.5 414 ms 516 ms 543 ms 0%
Product search 3.5 365 ms 469 ms 491 ms 0%
32 virtual users
Home page 3.41 1.74 s 1.99 s 2.17 s 0%
Shop page 3.41 2.29 s 2.6 s 2.71 s 0%
Category page 3.41 1.91 s 2.36 s 2.53 s 0%
Product page 3.41 1.72 s 1.91 s 1.98 s 0%
Product search 3.41 1.57 s 1.77 s 1.83 s 0%

Cart and checkout

Step Requests/s Median p95 p99 Error rate
8 virtual users
Add to cart 6.73 479 ms 605 ms 647 ms 0%
Cart 6.73 395 ms 505 ms 528 ms 0%
Checkout 6.73 310 ms 400 ms 421 ms 0%
32 virtual users
Add to cart 6.57 1.72 s 2.06 s 2.21 s 0%
Cart 6.57 1.69 s 1.94 s 2.03 s 0%
Checkout 6.57 1.37 s 1.61 s 1.72 s 0%

k6 with a fixed number of virtual users, 61 to 66 s per level. The error rate counts responses that weren't a 200.

Discourse

Signed in

Step Requests/s Median p95 p99 Error rate
1 virtual user
Topic page 1.95 233 ms 317 ms 477 ms 0%
Topic JSON 0.977 111 ms 146 ms 158 ms 0%
Home page 0.977 193 ms 327 ms 414 ms 0%
Categories page 0.497 228 ms 278 ms 437 ms 0%
User activity page 0.48 151 ms 239 ms 328 ms 0%
4 virtual users
Topic page 6.39 278 ms 387 ms 652 ms 0%
Topic JSON 3.2 145 ms 204 ms 236 ms 0%
Home page 3.19 250 ms 330 ms 515 ms 0%
Categories page 1.62 285 ms 380 ms 589 ms 0%
User activity page 1.6 183 ms 332 ms 451 ms 0%
16 virtual users
Topic page 6.15 1.06 s 1.29 s 1.58 s 0%
Topic JSON 3.09 850 ms 1.07 s 1.2 s 0%
Home page 3.11 1.02 s 1.26 s 1.55 s 0%
Categories page 1.54 1.1 s 1.36 s 1.6 s 0%
User activity page 1.55 875 ms 1.09 s 1.17 s 0%

Anonymous

Step Requests/s Median p95 p99 Error rate
16 virtual users
Topic page 148 28.5 ms 68.2 ms 95.6 ms 0%
Topic JSON 73.8 27.1 ms 62.8 ms 94 ms 0%
Home page 73.7 28.6 ms 63.9 ms 91.9 ms 0%
Categories page 36.9 28.4 ms 63.1 ms 94.4 ms 0%
User activity page 36.9 130 ms 180 ms 234 ms 0%

k6 with a fixed number of virtual users, 60 to 61 s per level, after a 21 s warmup. The error rate counts responses that weren't a 200. Discourse ran with 6 app workers and 1024 MB of Postgres shared buffers, what its installer picks for a 4 GB box. It added a 2 GB swapfile first, the way its installer does, and didn't need it. The container peaked at 2,614 MB.

Ghost

Readers

Step Requests/s Median p95 p99 Error rate
1 virtual user
Home page 8.99 30.1 ms 45 ms 54.2 ms 0%
Post page 22.4 16 ms 27.3 ms 33.2 ms 0%
Tag page 4.5 20.3 ms 32.6 ms 36.4 ms 0%
Author page 4.5 20.9 ms 31.7 ms 39.3 ms 0%
RSS feed 4.48 25.2 ms 41.3 ms 47.3 ms 0%
4 virtual users
Home page 12.1 88 ms 115 ms 130 ms 0%
Post page 30.2 60.7 ms 85.9 ms 98.1 ms 0%
Tag page 6.06 56.2 ms 77.4 ms 88.6 ms 0%
Author page 6.05 54.7 ms 74.9 ms 84.5 ms 0%
RSS feed 6.01 62.5 ms 84.5 ms 91 ms 0%
16 virtual users
Home page 11.7 347 ms 413 ms 441 ms 0%
Post page 29.4 284 ms 349 ms 380 ms 0%
Tag page 5.87 191 ms 247 ms 262 ms 0%
Author page 5.9 190 ms 249 ms 278 ms 0%
RSS feed 5.87 199 ms 249 ms 275 ms 0%

Staff in Ghost Admin

Step Requests/s Median p95 p99 Error rate
1 virtual user
Post list 4.83 29.4 ms 44.4 ms 49.6 ms 0%
Post editor 14.5 11.1 ms 18.2 ms 22.8 ms 0%
Saving a draft 14.5 30.4 ms 45.3 ms 54.2 ms 0%
Member list 4.83 35.3 ms 51.4 ms 63.6 ms 0%
4 virtual users
Post list 6.81 36.9 ms 48.1 ms 57 ms 0%
Post editor 20.5 18.2 ms 25.5 ms 32.9 ms 0%
Saving a draft 20.5 145 ms 183 ms 200 ms 0%
Member list 6.83 42.4 ms 56.1 ms 66.9 ms 0%

k6 with a fixed number of virtual users, 60 s per level, after a 20 s warmup. The error rate counts responses that weren't a 200. Ghost ran the way its ghost-docker setup runs it: one Node process, and MySQL with its stock 128 MB buffer pool. Memory peaked with Ghost at 309 MB and MySQL at 624 MB.

Wagtail

Visitors

Step Requests/s Median p95 p99 Error rate
1 virtual user
Home page 1.5 85.9 ms 115 ms 129 ms 0%
Blog index 1.48 289 ms 349 ms 420 ms 0%
Blog post 4.44 28.8 ms 46.9 ms 53.9 ms 0%
Bread page 4.44 32 ms 51.1 ms 57.9 ms 0%
Location page 1.48 30.6 ms 50.1 ms 54.1 ms 0%
Site search 1.48 52.9 ms 82.5 ms 99.7 ms 0%
4 virtual users
Home page 2.92 194 ms 279 ms 295 ms 0%
Blog index 2.97 649 ms 996 ms 1.07 s 0%
Blog post 8.81 52.4 ms 95.1 ms 121 ms 0%
Bread page 8.82 56.6 ms 91.7 ms 108 ms 0%
Location page 2.97 51.7 ms 87.1 ms 125 ms 0%
Site search 2.95 81.4 ms 169 ms 212 ms 0%
16 virtual users
Home page 3.1 599 ms 820 ms 883 ms 0%
Blog index 3.18 1.58 s 1.84 s 1.92 s 0%
Blog post 9.24 330 ms 535 ms 627 ms 0%
Bread page 9.37 335 ms 537 ms 611 ms 0%
Location page 3.13 326 ms 538 ms 630 ms 0%
Site search 3.1 388 ms 626 ms 680 ms 0%

Editors in the Wagtail admin

Step Requests/s Median p95 p99 Error rate
1 virtual user
Admin dashboard 0.676 91.1 ms 113 ms 123 ms 0%
Page explorer 1.38 342 ms 461 ms 509 ms 0%
Page editor 2.04 162 ms 203 ms 223 ms 0%
Image library 0.676 66.5 ms 102 ms 580 ms 0%
Admin page search 0.676 42.7 ms 75.7 ms 83.1 ms 0%
4 virtual users
Admin dashboard 1.09 233 ms 503 ms 555 ms 0%
Page explorer 2.23 794 ms 1.43 s 1.69 s 0%
Page editor 3.33 417 ms 849 ms 1.01 s 0%
Image library 1.11 216 ms 575 ms 719 ms 0%
Admin page search 1.09 98.8 ms 292 ms 614 ms 0%

k6 with a fixed number of virtual users, 60 to 61 s per level, after a 20 s warmup. The error rate counts responses that weren't a 200. Wagtail ran the way bakerydemo ships it: uWSGI with 2 workers of 4 threads each, whatever the box, and Postgres and Redis on stock settings. Memory peaked with Wagtail at 297 MB, Postgres at 188 MB, and Redis at 26 MB.

Advertised vs. measured

Advertised Measured
vCPU 3 3
CPU Shared vCPU 0% of busy CPU time stolen
RAM 4 GB 3.7 GiB
Storage 80 GB NVMe 75 GiB
Architecture x86_64 x86_64
Virtualization KVM KVM

Guests usually report slightly less RAM and disk than advertised because of kernel reservations and filesystem overhead.

Detected hardware

CPU model
AMD EPYC-Rome Processor
CPU cores
3
CPU frequency
2445 MHz
AES-NI
Yes
VM-x/AMD-V
No
RAM reported
3.7 GiB
Swap reported
0 MiB
Disk reported
75 GiB
Virtualization reported
KVM
CPU steal
0% of busy CPU time stolen
Distro
Ubuntu 24.04.4 LTS
Kernel
6.8.0-139-generic
Architecture
x86_64
IPv4 reachable
Yes
IPv6 reachable
Yes

Advertised at test time

Plan
CPX21
Product line
regular_purpose
Location
Hillsboro, OR
vCPU
3 shared vCPU
Architecture
x86_64
RAM
4 GB
Storage
80 GB NVMe
Virtualization
KVM
Network
Unknown
Transfer
2.2 TB
Port 25
On request
Monthly price at test time
$37.49
Hourly price at test time
$0.060
YABS version
v2026-09-20
Ran at
Oct 17, 2026, 06:35:44 UTC
Parser version
5

Run timeline

Queued
Oct 17, 2026, 06:35:00 UTC
First update from the server
Oct 17, 2026, 06:35:32 UTC
Completed
Oct 17, 2026, 08:06:44 UTC
Server deleted
Oct 17, 2026, 08:06:44 UTC
Cost of this run
$0.092
Watch party 0 watching

This party closed at . It's read only now.

Watch party

Loading the party