Hi there,
I want to benchmark performance of different OS'es on Raspberry Pi (Model B). Options are Raspbian and NetBSD... I will be using Raspberry Pi for GPIO or USB based automation and performance is important for the application (there will be constant data logging and serial communication through gpio or usb).
Is there a recommended method to benchmark different OSes, or is there any OS known to provide a better performance compared to others?
Thanks
-
DougieLawson
- Posts: 39304
- Joined: Sun Jun 16, 2013 11:19 pm
- Location: A small cave in deepest darkest Basingstoke, UK
- Contact: Website Twitter
Re: Benchmarking different OS'es
Stick with Raspbian, because that way you're getting the ideal performance complete with being in the largest group of users. It's the best supported and easiest to use OS.
If you find that you need to make things go faster then use sudo raspi-config to set overclocking.
If you find that you need to make things go faster then use sudo raspi-config to set overclocking.
Note: Any requirement to use a crystal ball or mind reading will result in me ignoring your question.
Criticising any questions is banned on this forum.
Any DMs sent on Twitter will be answered next month.
All non-medical doctors are on my foes list.
Criticising any questions is banned on this forum.
Any DMs sent on Twitter will be answered next month.
All non-medical doctors are on my foes list.
Re: Benchmarking different OS'es
If you can define what you mean by "better" you'll be half way to answering your own question.
I can't see how anyone can help. It's your application. You will have to run it in both environments with the data loads you expect and measure which one is "better".
I can't see how anyone can help. It's your application. You will have to run it in both environments with the data loads you expect and measure which one is "better".
Re: Benchmarking different OS'es
Overclocking should only be considered after you have attempted to improve the efficiency of your own codeDougieLawson wrote:If you find that you need to make things go faster then use sudo raspi-config to set overclocking.
PeterO
Discoverer of the PI2 XENON DEATH FLASH!
Interests: C,Python,PIC,Electronics,Ham Radio (G0DZB),1960s British Computers.
"The primary requirement (as we've always seen in your examples) is that the code is readable. " Dougie Lawson
Interests: C,Python,PIC,Electronics,Ham Radio (G0DZB),1960s British Computers.
"The primary requirement (as we've always seen in your examples) is that the code is readable. " Dougie Lawson
Re: Benchmarking different OS'es
fnoyanisi,
Given that you are wanting to create a complex system comprising GPIO, USB, serial, disk storage the recommended way to evaluate performance is to try you application on each of the possible platforms and monitor it's performance.
There are of course benchmarks for all kind of things. You can spend a lot of time running them and trying to interpret the results as they relate to your application. At the end you will still be none the wiser. And bottlenecks will appear out of the woodwork to thwart you app anyway.
I might ask, does it matter? Let's say operating system "A" is loaded at 10% when running your app. And operating system "B" is loaded at 60%. Obviously "B" is terrible, so slow.
Or is it? The app works right? Perhaps "B" is easier to use, has more support, has a bigger user community to get help from etc etc.
I suspect a lot of performance issues will come down to programming language choice, algorithm selection, implementation details. You should be concerned with those before worrying about the OS.
On a Pi, start with Raspbian. Look else where when you find your app really will not work as expected there.
Given that you are wanting to create a complex system comprising GPIO, USB, serial, disk storage the recommended way to evaluate performance is to try you application on each of the possible platforms and monitor it's performance.
There are of course benchmarks for all kind of things. You can spend a lot of time running them and trying to interpret the results as they relate to your application. At the end you will still be none the wiser. And bottlenecks will appear out of the woodwork to thwart you app anyway.
I might ask, does it matter? Let's say operating system "A" is loaded at 10% when running your app. And operating system "B" is loaded at 60%. Obviously "B" is terrible, so slow.
Or is it? The app works right? Perhaps "B" is easier to use, has more support, has a bigger user community to get help from etc etc.
I suspect a lot of performance issues will come down to programming language choice, algorithm selection, implementation details. You should be concerned with those before worrying about the OS.
On a Pi, start with Raspbian. Look else where when you find your app really will not work as expected there.
Memory in C++ is a leaky abstraction .