Do some apps show massive AMD/nvidia differences?

Mark R

Diamond Member
Oct 9, 1999
8,513
16
81
Just came across this strange thing at work.

We have a specialist app that does some lightweight 3D stuff. However, we have 2 configurations of workstation.

Primary workstations were equipped with 4k monitors; but lower-use workstations got dual 1440p monitors. Because of differences in connectivity between monitor configs, different cards were needed - the 4k monitors use dual DL-DVI connections. Whereas the 1440p monitors connect via DP.

For reasons of availability - the DL-DVI cards are nV Quadro FX380, and the DP cards are AMD Firepro V4900.

While I'm normally very happy with my FX380 card, I was testing a Firepro V4900, and I couldn't believe how slow it was in comparison - it was literally unusably laggy in comparison.

The application supplier said they were graphics card agnostic and just used openGL.

What surprised me is that the FX380 should completely outclass the Firepro 4900, when the AMD card is literally 2 GPU generations ahead in terms of "raw" performance.
 

Zxian

Senior member
May 26, 2011
579
0
0
Something else is up. Either the drivers are improperly installed, or the software isn't actually card agnostic.

The software we use at work (developed in house) was initially only tested on nVidia cards, since that's what all of our workstations have (Quadro 600). AMD and Intel support didn't come until we started needing to work from home, and some of us use AMD video cards. Performance of the graphics in our software falls in line with most benchmarks.
 

BrightCandle

Diamond Member
Mar 15, 2007
4,762
0
76
AMD is unfortunately well known for having shocking openGL drivers, has done for as long as I can remember. The saving grace for most gamers has been almost all games are DirectX. There is every possibility its just a performance bug that is causing it. Could be a problem with the driver install as well of course, that does occasionally happen with AMDs installer as well and a reinstall can often clear it. But if its happening consistently across all the machines then its probably actual performance. Could be a hardware difference that matters or it could just be problematic software.