Support Website Contact Support Blog

Laptop for pix4d

Hello

A few months ago I asked about maximum desktop computer for rapid processing of mosaic images. I received a very useful reply that gave excellent recommendation on a solid foundation that would later become our workstation for dominating image rendering…and dominate it does. I can render a 300 picture project in a about 20 minutes I typically have 5 projects per day to run and I’ll start on one and then the next and the next so I usually end up with 5 instances of pix4d running on 4 or 5 different projects and in about 45 minutes they are all done! Thank you again for your recommendations on this system!

Now we are looking for a laptop that we could use in the field to run pix4d on site. Unlike our desktop computer we are not looking for total world domination. Something that can render projects like mentioned above in a reasonable time (but just one at a time I wont be running 5 instances of pix4d) we would like to be around the 4,000 dollar mark budget wise. Is there anything out there in this price range that would be a suitable system for rendering a 300 picture mosaic that is of good quality (stable platform, 100% image calibration, good control points, good image match, low focal relative difference)…

I am skeptical that a lap top has the heat dissipation to pack the horsepower needed for this kind of work. Our desktop work station has a gtx titan z video card and a noctua dual fan heat sink for the processor these two items alone are larger than a laptop.

Please let me know what you might recommend for this type of work and how the processing time might compare to our desktop system.

Thanks

-Nick

I think your best bet is to look into what 3D animators are using for portable work stations. Programs like Blender 3D have similar, if not more extreme, hardware requirements to Pix4D. If I remember correctly, GPU is not actually used for much of Pix4D’s processing so you will want to be more conscious of CPU speed and ventilation.

This link has several options that will work for you: http://blog.render.st/best-notebooks-for-3d-rendering-part-2-recommended-configurations/

  • Jacob

Pix4d requires a balanced system free of bottle necks and can utilize nvida cuda cores (processor, gpu, and ram should all be matched)

Could you provide an approximate render time for a 300 image mosaic that as mentioned initially has well calibrated images with low error using a laptop like the Sager NP9377-S from the link you provided?

Thanks

-Nick

Hi Nick,

As I am not employed by Pix4D I don’t have enough experience to estimate a processing time on a given computer. Your current desktop is a beast of a computer and can process much faster than my system can. My system is actually similar in specification to the Sager NP9377-S so let me give you an idea of my processing
Specs:
-CPU: Intel® Core™ i7-4770K CPU @ 3.50GHz, 3501 Mhz, 4 Core(s), 8 Logical Processor(s)
-GPU: Nvidia GTX Titan X 12GB GDDR5
-RAM: 32GB
-Motherboard: Z97MX-Gaming 5
-Memory: 250GB SSD w/ 2TB harddisk drive (Pix4D folder and files stored on SSD, images stored on HDD)

Take a look at this quality report http://1drv.ms/1Vbd2GB for an idea of my processing time given the options I select. My CPU and/or cooling seems to be the major bottleneck. I never max out RAM, but it always runs @ 100% CPU usage.

Jacob

If I read that right It looks like it took about 15min to get past step 1 where it generates the first quality report with a pretty much ideal data set

Thats actually pretty helpful! Thank you.

Did you fly that with a multi rotor?

Nick,

I’d be interested to know what was recommended for your desktop configuration.  I’m using a laptop that takes about 6 hours to run with 500 images.

Thanks,

Joe Zamudio

https://community.pix4d.com/t/4384-Maximum-computer-for-time-reduction- 

 

Im using super computer recommendation with titan z GPU

Dell 6800 precision

32g ram

k5100m gpu

i7 4940mx

OWC SSDs

Dock stations and 27" displays for office and work trailer

not cheap but gives me peace of mind and quick results in the for data confidence.

FWIW