abonneren

How fast is the Raspberry Pi 3? Benchmarking with Sysbench

The Raspberry Pi 3 has been released, and we've been able to get our hands on the mini-computer for a time now. Besides a built-in wifi and bluetooth-module, not much is new about about the Pi, though it does have a new processor. We took Sysbench to measure the speed of the cpu to see if it really is faster than earlier models.

The Raspberry Pi 3 is indeed more powerful than its predecessors. The third generation mini-computer has a Broadcom BCM2837-cpu, as opposed to the BCM 2836-soc in the 2 Model B+. That means the new soc has a clockspeed of 1,2 GHz, as opposed to 900 MHz in its predecessor.

RAM stays the same, at 1 GB.

We ran Sysbench on 3 different devices: The Raspberry Pi 3 against a Raspberry Pi 2 Model B+, and we threw in an original Raspberry Pi 1 for good measure.
 

Benchmark with Sysbench

You can see the Pi 3 is about 40% faster than its predecessor. These are the results for the Raspberry Pi 3:

Running the test with following options:
Number of threads: 4

Doing CPU performance benchmark

Threads started!
Done.

Maximum prime number checked in CPU test: 10000


Test execution summary:
    total time:                          45.8901s
    total number of events:              10000
    total time taken by event execution: 183.4899
    per-request statistics:
         min:                                 18.21ms
         avg:                                 18.35ms
         max:                                 38.38ms
         approx.  95 percentile:              18.48ms

Threads fairness:
    events (avg/stddev):           2500.0000/17.76
    execution time (avg/stddev):   45.8725/0.00
 



These are the results for the Raspberry Pi 2 Model B:

Running the test with following options:
Number of threads: 4

Doing CPU performance benchmark

Threads started!
Done.

Maximum prime number checked in CPU test: 10000


Test execution summary:
    total time:                          76.1168s
    total number of events:              10000
    total time taken by event execution: 304.4156
    per-request statistics:
         min:                                 29.65ms
         avg:                                 30.44ms
         max:                                 63.32ms
         approx.  95 percentile:              34.97ms

Threads fairness:
    events (avg/stddev):           2500.0000/7.38
    execution time (avg/stddev):   76.1039/0.01
 



And, as promised, the results for the originele Raspberry Pi:

    Running the test with following options:
    Number of threads: 4
     
    Doing CPU performance benchmark
     
    Threads started!
    Done.
     
    Maximum prime number checked in CPU test: 10000
     
     
    Test execution summary:
        total time:                          523.1061s
        total number of events:              10000
        total time taken by event execution: 2091.9841
        per-request statistics:
             min:                                162.66ms
             avg:                                209.20ms
             max:                                252.29ms
             approx.  95 percentile:             232.33ms
     
    Threads fairness:
        events (avg/stddev):           2500.0000/1.22
        execution time (avg/stddev):   522.9960/0.04

Geschreven door: Tijs Hofmans op

Category: Nieuws, Domotica

Tags: Domotica, Raspberry Pi 3, benchmark, sysbench, Raspberry Pi, Raspberry

Nieuws headlines

Laatste reactie