Perf. differences between i7-7820 laptop and xeon workstations

PIX4Dmapper PIX4Dmapper Questions/Troubleshooting
1

Dear Sirs,

I need to upgrade the workstation used by future pix4d operator and I have a couple of questions.

Current hardware:

ZBOOK 15

i7-4800MQ 2.70

24Gb RAM

Nvidia QUADRO K2100

 

With this configuration latest job (about 800 images) was completed in about 5 hours

 

The person will frequently travel but we’re evaluating to move to a desktop workstation in order to get better and faster resutls. Here the proposed HP Workstation Z640 E5-2630v4 with 64Gb RAM (portability will be solved with an additional laptop)

GPU we have in the list a K2200 or M4000 or maybe a TITAN

Approx how much faster we could expect to get results compared to a new laptop configuration:

ThinkPad P51 20HH-001R i7-7820HQ 2x8GB NVME SSD 512GB (upgraded to 32/64Gb)

If we proceed with a workstations with options could give us best benefits?

  • additional xeon processor

  • got the faster GPU (M4000?)

  • others?

  • RAID drives

Both configuration will be upgraded to at least 1TB SSD disk

 

Latest question: if we downgrade to a slower xeon proc. (16…) we lost a lot in term of performances (same RAM and GPU)?

 

Thanks in advance.

 

 

2

Xeon processor is definitely not faster and could be slower.  And a GeForce card is going to be the best for the money.

More RAM changes the clusters it uses in point cloud creation so in addition to speed, the accuracy might get tweaked.  Different hardware tweaks the accuracy but may not be noticeable.

I doubt anybody here can quote you specifics but I have been on a myriad of different hardware to find more speed and it just isn’t out there.

CPU: Intel(R) Core™ i7-5960X CPU @ 3.00GHz
RAM: 64GB
GPU: NVIDIA GeForce GTX 1060 6GB (Driver: 22.21.13.8205)

Raw GHz, RAM and GPU RAM are what you want to maximize.  And yes a 512GB SSD rounds out the max performance criteria.

3

And maybe Linux to optimize project, systems and resource management, one can dream.

Thank you very much Adam, and all others on the quest for hardware optimization…Thanks!!!