sysbench results / cpufreq

jimm0thy

I was interested in checking the performance against my Pi 3b and also wanted to know the speed the cpu was running at. I needed to load scpi-cpufreq module to get any information out of cpufreq which tells me the cpu on the La Frite is running at 1ghz idle and 1.5ghz load.
 (interesting since the info for the board says 1.2ghz max speed, perhaps cpufreq is wrong I dunno)

 My Pi 3b is overclocked to 1.3ghz.
The La Frite is running the Debian-headless-with-mali image , and the pi is running RetroPie 4.5

The La Frite completed the test in 6.18s

The Pi 3b completed the test in 85.73s

Ingmar

Nice.
If only I could get this one to run at least one image

FransM

No idea on the clock frequency but sysbench is not a great tool to compare frite and rpi.

See https://www.cnx-software.com/2019/04/30/giggle-score-odroid-n2-best-value-raspberry-pi-zero-worst-value/#comment-562491

VCR0

Ingmar said:

Nice.
If only I could get this one to run at least one image

LOL, that's what i'm at now.  Great performance from command line, abysmal performance in Desktop Environment.

marcofranchimoretti

SysBench Update.

LeFritte libre-computer-aml-s805x-ac-debian-buster-headless-mali-4.19.64+

LePatato libre-computer-aml-s905x-cc-debian-stretch-headless-mali-4.19.55+

sudo apt-get install curl

Install Sysbench: curl -s https://packagecloud.io/install/repositories/akopytov/sysbench/script.deb.sh | sudo bash

sudo apt-get install sysbench sysbench --test=cpu --threads=4 run

LeFritte sysbench 1.0.17 (using bundled LuaJIT 2.1.0-beta2) Running the test with following options: Number of threads: 4 Initializing random number generator from current time Prime numbers limit: 10000 Initializing worker threads... Threads started!

CPU speed: events per second: 2395.84 General statistics: total time: 10.0014s total number of events: 23976 Latency (ms): min: 1.66 avg: 1.67 max: 1.84 95th percentile: 1.67 sum: 39985.19 Threads fairness: events (avg/stddev): 5994.0000/0.71 execution time (avg/stddev): 9.9963/0.00

sysbench --test=memory --threads=4 run sysbench 1.0.17 (using bundled LuaJIT 2.1.0-beta2) Running the test with following options: Number of threads: 4 Initializing random number generator from current time Running memory speed test with the following options: block size: 1KiB total size: 102400MiB operation: write scope: global Initializing worker threads... Threads started! Total operations: 37692608 (3766945.26 per second) 36809.19 MiB transferred (3678.66 MiB/sec) General statistics: total time: 10.0002s total number of events: 37692608 Latency (ms): min: 0.00 avg: 0.00 max: 4.18 95th percentile: 0.00 sum: 17605.81 Threads fairness: events (avg/stddev): 9423152.0000/19998.58 execution time (avg/stddev): 4.4015/0.01

LePatato sysbench 1.0.14 (using bundled LuaJIT 2.1.0-beta2) Running the test with following options: Number of threads: 4 Initializing random number generator from current time Prime numbers limit: 10000 Initializing worker threads... Threads started! CPU speed: events per second: 2391.80 General statistics: total time: 10.0012s total number of events: 23935 Latency (ms): min: 1.67 avg: 1.67 max: 2.22 95th percentile: 1.67 sum: 39986.32 Threads fairness: events (avg/stddev): 5983.7500/0.43 execution time (avg/stddev): 9.9966/0.00

sysbench --test=memory --threads=4 run sysbench 1.0.14 (using bundled LuaJIT 2.1.0-beta2) Running the test with following options: Number of threads: 4 Initializing random number generator from current time Running memory speed test with the following options: block size: 1KiB total size: 102400MiB operation: write scope: global Initializing worker threads... Threads started! Total operations: 36677584 (3665553.12 per second) 35817.95 MiB transferred (3579.64 MiB/sec) General statistics: total time: 10.0001s total number of events: 36677584 Latency (ms): min: 0.00 avg: 0.00 max: 0.18 95th percentile: 0.00 sum: 17349.57 Threads fairness: events (avg/stddev): 9169396.0000/1009.07 execution time (avg/stddev): 4.3374/0.00

loverpi

Desktop performance is abysmal because of 3D driver and X11 stack. Raspberry Pi hacks around this by using X11's fbturbo. fbturbo performs just as fast on La Frite but it loses 3D support. The proper alternative is wayland but LXQT is only in beta right now. We will work on a wayland image for better desktop performance once 5.3 is tagged. Linux 5.4 will be the LTS and we will have accelerated video decode working on that release within a wayland desktop compositor.