[Marius]

What are the system requirements to run TopoDOT?

[Randy Allen]

Hi Marius,

Check out this link: https://new.certainty3d.com/blog/technotes/ for some great Tech Notes.

V8i: https://new.certainty3d.com/blog/topodot...uirements/

Connect: https://new.certainty3d.com/blog/topodot...uirements/

[Marius]

Hi Marius,

Check out this link: https://new.certainty3d.com/blog/technotes/ for some great Tech Notes.

V8i: https://new.certainty3d.com/blog/topodot...uirements/

Connect: https://new.certainty3d.com/blog/topodot...uirements/

Hello Randy, 

Thanks for your reply. 

Did you test TopoDot - for microstation connect - on a PC that has has more than "32GB DDR4-2113" & a better graphics card >"Video Memory: 4+ GB (NVIDIA GeForce GTX 1060)". 

If so, does TopoDot operate faster with large amounts of data (100Gb - 200Gb of point cloud data). Does it reduce the processing times?

[Randy Allen]

Hi Marius,

Check out this link: https://new.certainty3d.com/blog/technotes/ for some great Tech Notes.

V8i: https://new.certainty3d.com/blog/topodot...uirements/

Connect: https://new.certainty3d.com/blog/topodot...uirements/

Hello Randy, 

Thanks for your reply. 

Did you test TopoDot - for microstation connect - on a PC that has has more than "32GB DDR4-2113" & a better graphics card >"Video Memory: 4+ GB (NVIDIA GeForce GTX 1060)". 

If so, does TopoDot operate faster with large amounts of data (100Gb - 200Gb of point cloud data). Does it reduce the processing times?

Hi Marius,

I have not. I'm getting ready to update to a desktop, so I'll keep you posted. 

My current laptop configuration can process (tiling & classification) about 50 gigs per hour. 
Intel® Core™ i7-4800MQ CPU @ 2.70GHz
32 GB RAM
Intel® HD Graphics 4600

So, yeah....I'm looking forward to the larger setup. 

[David Terneus]

Hello Randy, 

Thanks for your reply. 

Did you test TopoDot - for microstation connect - on a PC that has has more than "32GB DDR4-2113" & a better graphics card >"Video Memory: 4+ GB (NVIDIA GeForce GTX 1060)". 

If so, does TopoDot operate faster with large amounts of data (100Gb - 200Gb of point cloud data). Does it reduce the processing times?

Marius, 

It depends on what type of processing you're doing. If you're extracting a topography model on a road corridor, for example, it's beneficial to only work on a couple hundred million points at a time (and our tools make it easy to load a couple 'tiles' at a time) to have the tools work as quick as possible.

[Randy Allen]

Hello Randy, 

Thanks for your reply. 

Did you test TopoDot - for microstation connect - on a PC that has has more than "32GB DDR4-2113" & a better graphics card >"Video Memory: 4+ GB (NVIDIA GeForce GTX 1060)". 

If so, does TopoDot operate faster with large amounts of data (100Gb - 200Gb of point cloud data). Does it reduce the processing times?

Marius, 

It depends on what type of processing you're doing. If you're extracting a topography model on a road corridor, for example, it's beneficial to only work on a couple hundred million points at a time (and our tools make it easy to load a couple 'tiles' at a time) to have the tools work as quick as possible.

I took the question as processing batch jobs .... such as tiling (parsing) and classification. 

I would agree if doing topo work, to minimize the work area to a couple tiles....I can generally load about 6 gigs of data on my current laptop without lag.

[Marius]

Thanks for your replies. 

Randy, that is exactly what I had in my mind (batch processing) but lacked the words... 

I'm looking to speed up the point cloud deviation / ground classification / optimization without reducing the number of points or limiting it.   

Ex: if a point cloud is being optimized and converted to .dot, when exported to .las the file is loosing some of the points. When using TopoDot it doesn't matter that much, but in other software the difference can be spotted straight away. 

So can the batch processing of data be faster if the computer has more Ram for example?
Or the High end specifications is the limit, meaning having more "power" in your PC doesn't actually help.

[mterneus]

Ex: if a point cloud is being optimized and converted to .dot, when exported to .las the file is loosing some of the points. 

Hello Marius,

What do you mean by this? You should not be losing any of the points on export.


As for your other questions,

With regards to point cloud processing:

Large generational jumps will make a big difference.  Examples:
HDD --> SSD\NVMe
SD Ram --> DDR4
Single Core --> Multiple cores
Old Architecture ---> Newer Architecture... what I mean by this is the newer Core I7 processors are much more efficient than the first generation Core I7 processors regardless of clock speed.


I would suggest investing in a fast NVMe SSD drive (Example: https://smile.amazon.com/Samsung-970-PRO...way&sr=8-5)

This would speed up the loading \ exporting of data which I think is the biggest performance boost you will get if coming from an HDD. 


Having a lot of data loaded at one time (RAM) would allow you to process bigger tiles at one time, however there are restrictions on certain tools that would limit how big of an area you can do at one time....
For example, the point cloud deviation tool has limitations on how large of an image we are able to create... so if you are processing a tile larger than that limitation, you will not be able to run the tool.

The tools you are looking to run will handle loading \ unloading data tiles for you and it should not really be much if any faster to run it on the entire data set at once.


Video memory is not used too much except for MicroStation rendering using textures and things.. Having more of that will not help too much.  Typically faster clock speeds should be what you need to shoot for... but again.. unless making a big generational jump the difference will be marginal..  The same goes for processors.

TopoDOT will utilize multiple cores, however the number of cores that TopoDOT utilizes scales depending on how much data is being processed.  Simple operations such as taking a cross section will probably not use more than 1 or 2 cores, but tools like classification of large amounts of data may utilize all of the cores.

[Marius]

Ex: if a point cloud is being optimized and converted to .dot, when exported to .las the file is loosing some of the points. 

What do you mean by this? You should not be losing any of the points on export.

This was just an example, I'm still just a junior TopoDot user and am aware that I might not understand or have the right boxes ticked in TopoDot. Nothing serious to worry about. Also it's possible that it's more difficult to describe in words my experience when it is a matter of show and tell. I hope that this is not perceived in any other way...  
Maybe this is a subject that would be relevant in a different topic. 

What I've been looking for, you've described in the last post, in terms of PC specifications. 

A big Thank you!
Marius

---

This are my PC spec.:

CPU:     Intel® Xeon® CPU E5-2620 v3 @ 2.40GHz
RAM:    32 Gb DDR4
HDD:    Samsung SSD 850 EVO 500GB
Nvidia Quadro K620.

Would it make any difference if we were to configure a new PC based on the High End Spec: https://new.certainty3d.com/blog/topodot...uirements/?

Or is it a matter of adding more RAM and a fast NVMe SSD drive? 

[mterneus]

Hi Marius,

The technote is a little old and could use a little updating.

Your processor is a bit old and has a slow clock speed. 
I think you may benefit a good bit by grabbing a new processor with a higher clock speed.. Ex: Core I9-9900k (~500$)


https://cpu.userbenchmark.com/Compare/In...5895vs4028

I do not have much experience with Quadro cards to give you advice on that.

More Ram will allow you to pull in more data, so if that is what you are looking for, you can add more.