Want to support development? Maybe get a nice server to do your own development on? Check out BuyVM or Vultr using our affiliate links! New users get $50 of free credit with Vultr, and BuyVM offers affordable servers with affordable, true DDoS filtering. You can also support us on Patreon.
Description: some SOC boards benchmarks
Submitted on October 29, 2016 at 06:32 AM

Section 1 (Text)

Name                              Benchmark details                                score core Type
Beaglebone Black                  http://browser.primatelabs.com/geekbench2/2356217  248 1    Cortex-A8
Raspberry Pi 1 Model B @ 700 MHz  http://browser.primatelabs.com/geekbench2/2581041  250 1    ARM1176JZF-S
Raspberry Pi 1 Model B @ 1000 MHz http://browser.primatelabs.com/geekbench2/2581045  373 1    ARM1176JZF-S
Banana PI BPI-M1 [Debian 8]       http://browser.primatelabs.com/geekbench2/2581046 1004 2    Cortex-A7
Raspberry Pi 2 model B @ 700MHz   http://browser.primatelabs.com/geekbench2/2545902 1042 4    Cortex-A7
Intel Atom N450 @ 1.67 GHz        http://browser.primatelabs.com/geekbench2/2581101 1169 1/2  [1 core, 2 threads]
Raspberry Pi 2 model B @  900 MHz http://browser.primatelabs.com/geekbench2/2545903 1313 4    Cortex-A7 
Raspberry Pi 2 model B @ 1000 MHz http://browser.primatelabs.com/geekbench2/2545907 1450 4    Cortex-A7
Banana PI BPI-M2                  http://browser.primatelabs.com/geekbench2/2569045 1556 4    Cortex-A7
i.MX-6 1.20 GHz 4 core            http://browser.primatelabs.com/geekbench2/2540086 1833 4    
ODROID-C1 1.5 GHz                 http://browser.primatelabs.com/geekbench2/2532230 1924 4    Cortex-A5
ODROID-C1 1.73 GHz                http://browser.primatelabs.com/geekbench2/2565193 2026 4    Cortex-A5
Raspberry Pi 3 Model B @1200 MHz  http://browser.primatelabs.com/geekbench2/2596704 2128 4    Cortex-A53
Odroid U3                         http://browser.primatelabs.com/geekbench2/2422084 2749 4    Cortex-A9
Odroid-C2 Amlogic S905 @2000 MHz                                                   ~3300 4    Cortex-A53
NVIDIA Jetson TK1                 http://browser.primatelabs.com/geekbench2/2465601 3425 4    Cortex-A15
AMD Athlon II X2 250e @ 3.00 GHz  http://browser.primatelabs.com/geekbench2/2581036 3698 2 
ODROID-XU3                        http://browser.primatelabs.com/geekbench2/2562801 5037 4+4  Cortex-A15+Cortex-A7
ODROID-XU4 [Ubuntu 15.04]         http://browser.primatelabs.com/geekbench2/2581032 5325 4+4  Cortex-A15+Cortex-A7
 
The GeekBench2 tests are:
Integer
  Blowfish
  Text Compress
  Text Decompress
  Image Compress
  Image Decompress
  Lua
 
 
Floating Point
  Mandelbrot
  Dot Product
  LU Decomposition
  Primality Test
  Sharpen Image
  Blur Image
 
Memory
  Read Sequential
  Write Sequential
  Stdlib Allocate
  Stdlib Write
  Stdlib Copy
 
Stream (see https://www.cs.virginia.edu/stream/ref.html#counting )
  Stream Copy
  Stream Scale
  Stream Add
  Stream Triad
 
 
Geekbench only have an ARM Linux tool for one version of their GeekBench 2, and none in their latest and greatest GeekBench 3 (The non trial version of their tool sells for $13, so the market for an ARM Linux benchmark tool is tiny, they do have Android and iOS covered)
 
Here is how I ran the tests on some ARM Linux machines that I own:
$ wget http://cdn.primatelabs.com/geekbench-2.4.2-LinuxARM.tar.gz
$ tar xvfz geekbench-2.4.2-LinuxARM.tar.gz
$ cd dist/Geekbench-2.4.2-LinuxARM
$ ./geekbench --benchmark
(when the benchmark finishes running they give a URL to where they have uploaded the results on their website)
 
Other trial versions can be downloaded from the following URL's:
http://cdn.primatelabs.com/Geekbench-2.4.2-Linux.tar.gz
http://cdn.primatelabs.com/Geekbench-2.4.2-WindowsSetup.exe
http://cdn.primatelabs.com/geekbench-2.4.2-LinuxARM.tar.gz
http://cdn.primatelabs.com/Geekbench-3.3.2-Linux.tar.gz
http://cdn.primatelabs.com/Geekbench-3.3.2-WindowsSetup.exe
 
The Trial versions only test the 32 bit instruction set and the results can only be viewed online. And most version fails at detecting multiple cores and threats them as threading opportunities to parallel the testing. The testing only looks at CPU performance, so even though a NVIDIA Jetson TK1 has a good CPU and an amazing GPU, this is not reflected in the results.