Entry level graphics card for point cloud processing
- Jason Warren
- Administrator
- Posts: 4224
- Joined: Thu Aug 16, 2007 9:21 am
- 16
- Full Name: Jason Warren
- Company Details: Laser Scanning Forum Ltd
- Company Position Title: Co-Founder
- Country: UK
- Skype Name: jason_warren
- Linkedin Profile: No
- Location: Retford, UK
- Has thanked: 443 times
- Been thanked: 246 times
- Contact:
Entry level graphics card for point cloud processing
I did use Nvidia based cards myself, having had a couple of Quadro`s that overheated and burnt out I`m now using an AMD R7 card.
I`m always looking to speed up processing times and would like to hear anyone`s comments on the AMD R9 290X.
Specification
Radeon R9 290X
- Backplate included: Yes
- GPU: Hawaii XT
- Litecoin Hash Rate: 850-1000
- Stream Processors: 2816
- ROPS: 64
- Core Speed: 1040MHz
- Memory Speed: 5000Mhz
- Memory interface: 512-Bit
- Memory capacity: 8192MB GDDR5
- Cooling: MSI Twin FrozR Gaming Cooling
- PCI-Express X16 lane required
- 600W or greater PSU required
- Power Connectors: 8-pin + 6-pin required
- Display Outputs: 2x Dual Link DVI, 1x HDMI, 1x DisplayPort
- Warranty: 3yr
Features
- DirectX 11.2 Support
- Open GL 4.2 Support
- Dolby TrueHD and DTS-HD Master Audio support
- Crossfire Support upto 4-way: Native software crossfire (No bridge required)
- Gaming @ 4k Resolutions (UltraHD)
- Synergy with next-gen game consoles (Mantle)
- Graphics Core Next (GCN) architecture - optimized from previous generation products
https://www.overclockers.co.uk/showprod ... =GX-269-MS
The spec looks great, and affordable at around £300.00, but at the end of the day can point cloud processing software make good use of all the tech on these cards?
You do not have the required permissions to view the files attached to this post.
Jason Warren
Co_Founder
Dedicated to 3D Laser Scanning
LaserScanningForum
Co_Founder
Dedicated to 3D Laser Scanning
LaserScanningForum
-
- V.I.P Member
- Posts: 128
- Joined: Sun Feb 02, 2014 10:31 pm
- 10
- Full Name: David McNamara
- Company Details: Freelance
- Company Position Title: Research
- Country: United States
- Linkedin Profile: Yes
-
- I have made 50-60 posts
- Posts: 54
- Joined: Thu Jan 16, 2014 7:35 am
- 10
- Full Name: Marcin
- Company Details: Not Available
- Company Position Title: 3D Project Manager
- Country: Poland
- Linkedin Profile: No
- Location: Cracow, Poland
Re: Entry level graphics card for point cloud processing
David draw attention at right matter, that everything is depend on which software you want to use with this hardware. I am working with Bentley microstation, and I didn't seen any difference in performance using dedicated and integrated graphic (intel 4600) cards. The most important issue in this case is connected with CPU power.
My observation was confirmed by Bentley support team once I've asked them for information in this issue.
"We’ve compared timings of various point cloud data sets on different configurations, and it doesn’t look like the choice of graphics hardware makes much difference in the overall performance. The most important criteria are a good fast CPU and adequate system memory.
In general, most architectures with on-board graphics have weaker graphics capability and don’t perform as well as separate graphics cards for most MicroStation applications. However, the tight integration between CPU and GPU of these architectures, where both can access the same memory, is advantageous for point cloud display, where we have large amounts of data but relatively little computation. So, while higher end AMD or NVIDIA graphics hardware is generally best for most MicroStation applications, the Intel on-board graphics (for example, an i5 or i7 core with Intel HD4000 or better graphics) performs very well for point cloud display"
My observation was confirmed by Bentley support team once I've asked them for information in this issue.
"We’ve compared timings of various point cloud data sets on different configurations, and it doesn’t look like the choice of graphics hardware makes much difference in the overall performance. The most important criteria are a good fast CPU and adequate system memory.
In general, most architectures with on-board graphics have weaker graphics capability and don’t perform as well as separate graphics cards for most MicroStation applications. However, the tight integration between CPU and GPU of these architectures, where both can access the same memory, is advantageous for point cloud display, where we have large amounts of data but relatively little computation. So, while higher end AMD or NVIDIA graphics hardware is generally best for most MicroStation applications, the Intel on-board graphics (for example, an i5 or i7 core with Intel HD4000 or better graphics) performs very well for point cloud display"
- JohnT2
- I have made 40-50 posts
- Posts: 45
- Joined: Mon Feb 17, 2014 6:33 pm
- 10
- Full Name: John Tocci Jr
- Company Details: PCL Construction
- Company Position Title: VDC Manager
- Country: United States
- Linkedin Profile: Yes
- Been thanked: 1 time
Re: Entry level graphics card for point cloud processing
It's interesting that you had a bad experience with the Quadro cards... In 2012, we bought 4 Dell M6600s with the FirePro M8900... the machines worked great, for a time.
One by one, they have all gone down with flat-out video card failures while their Quadro counterparts (purchased both before and after) work just fine.
If we had all sorts of time and money, we would certainly test every promising card, but unfortunately since our M6600 die-off, we've had to avoid the AMD video cards out of prudence.
As an aside: Autodesk + Nvidia always seem to play nicer than Autodesk + AMD...
...but like anything on the internet and based on a statistically small selection set: YMMV
One by one, they have all gone down with flat-out video card failures while their Quadro counterparts (purchased both before and after) work just fine.
If we had all sorts of time and money, we would certainly test every promising card, but unfortunately since our M6600 die-off, we've had to avoid the AMD video cards out of prudence.
As an aside: Autodesk + Nvidia always seem to play nicer than Autodesk + AMD...
...but like anything on the internet and based on a statistically small selection set: YMMV
-
- V.I.P Member
- Posts: 128
- Joined: Sun Feb 02, 2014 10:31 pm
- 10
- Full Name: David McNamara
- Company Details: Freelance
- Company Position Title: Research
- Country: United States
- Linkedin Profile: Yes
Re: Entry level graphics card for point cloud processing
One other thing to note is that some features/technologies are not present in Nvidia cards that are in Quadro/Firepro cards, the biggest one bring quad buffered openGL stereo technology. You might not need this depending on the applications your running, but if you do you will need to run at least one quadro card to enable it.
I run a single Quadro K2000 and 3 GTX780 cards in order to enable features I need and still get massive amounts of render power for lower cost.
I run a single Quadro K2000 and 3 GTX780 cards in order to enable features I need and still get massive amounts of render power for lower cost.
-
- V.I.P Member
- Posts: 201
- Joined: Sun Oct 27, 2013 6:50 pm
- 10
- Full Name: Arash Yaghoubi
- Company Details: Hypsometric
- Company Position Title: Director of Cartography
- Country: USA
- Linkedin Profile: No
- Been thanked: 3 times
Re: Entry level graphics card for point cloud processing
SSDs seem to be determining pointcloud performance recently . But as dense meshes become the norm I'm hoping our investments payoff.