Ask Your Question
0

Laptop: Minimum System Requirement for OpenCV

asked 2017-12-07 01:03:59 -0600

Santhosh1 gravatar image

updated 2017-12-07 01:14:47 -0600

Hi guys,

Couldn't find anything solid regards with the kind of solution I am after online, so had to post this here.

Please note I can't build a desktop which is definitely much better for such work.

I mainly work on OpenCV. Currently doing good with python. C++ is an option. Learning cython in hope of trying to run the code faster.

Can't afford an Asus ROG Zephyrus GX501 or anything equally good right now, which can also be used for training NN due to GTX 1080's CUDA core count of 2560

I wanted to know what is the minimum system requirement to run OpenCV + Python or OpenCV + C++ Official site doesn't actual mention anything with regards to the actual minimum requirement.

I would like to know what could be my best bet, please suggestion a few laptops which can work good for my present situation?

Thanks,

edit retag flag offensive close merge delete

4 answers

Sort by ยป oldest newest most voted
3

answered 2017-12-07 06:11:47 -0600

kbarni gravatar image

OpenCV will run on anything. Even on a Raspberry Pi 1 (700MHz single core CPU, 256Mb memory). As long as it can run Windows or Linux, you're mostly OK.

Then, everything depends on your processing needs, and you didn't tell what do you want to do. There are some very fast algorithms, but other need lots of computing power. Same for the memory: processing a 640x480 pixel image from a webcam needs little memory, while processing a 4K video will fill your RAM quickly.

Coding on GPU is quite hard. Your code will run mostly on CPU, so the video card is irrelevant. The GPU might be important for other type of tasks, for example if you want to train deep neural networks - but then I would recommend a desktop computer, as it's more modular, powerful and much cheaper.

edit flag offensive delete link more

Comments

Good response. I'll add this:

OpenCV is a set of tools and various optimized ways to solve small problems. Optimization is mostly in how you use what's provided:

  • Feed OpenCV just enough input data to solve the end problem, no more
  • Choose between OpenCV function alternatives intelligently

All languages call the same precompiled OpenCV and numpy library functions. For many problems, where most of the time is spent in the library calls, Python runs just about as fast as C++, within a few percent. Python tends to be easier to write, and no fancy compilers are needed.

If the hardware just isn't up to the job (doesn't support more optimal ways, or is just too much data for the available time), compiling won't help much.

opalmirror gravatar imageopalmirror ( 2017-12-08 14:43:56 -0600 )edit
0

answered 2019-03-28 06:11:50 -0600

opencv-3.4.0 needs to compile atleast 1GB RAM with 2GB swap memory compulsory cause less than this configuration opencv will not able to compile on that system to generate object code. If u download binary version then it will run on minimum conf than this.

edit flag offensive delete link more
0

answered 2017-12-11 01:28:52 -0600

Santhosh1 gravatar image

@ya_ocv_user and @kbarni, really good feed back to have on the lowest possible hardware on which opencv works.

I did a bit of research, since my laptop(Core2Duo) is really old. A laptop with a latest 8th Gen of Intel Core i7-8550U seems to a balance between performance, portability and battery life. I feel its worth spending on it

edit flag offensive delete link more
0

answered 2017-12-07 01:16:28 -0600

Nothing special. I use not very fast laptop with 2Gb memory, Windows 8, and Visual Studio 2015. It works without visible restrains.

edit flag offensive delete link more

Question Tools

2 followers

Stats

Asked: 2017-12-07 01:03:59 -0600

Seen: 22,831 times

Last updated: Dec 11 '17