NVIDIA's Current Linux Driver Is Hungry For vRAM This Holiday

Written by Michael Larabel in NVIDIA on 30 November 2017 at 05:32 PM EST. 13 Comments
NVIDIA
With a NVIDIA Linux developer having confirmed a current driver performance regression affecting driver releases since the 378 series and not being worked around until the yet-to-be-released 390.xx beta driver, I decided to carry out some tests.

The regression as outlined in the aforelinked article is causing a performance drop in some Linux games with the NVIDIA driver releases after the 378 series. The issue is related to memory management changes but the precise issue is still being analyzed at NVIDIA.
gtx-1060-nv-reg-testing

I started off my testing with a GeForce GTX 1070 and comparing the "good" 375.82 driver to the latest 387.34 driver release. That's the first card tested so far with hours of benchmarking for this comparison due to running a wide-range of tests due to the lack of clarity over this Linux driver issue.
gtx-1060-nv-reg-testing

gtx-1060-nv-reg-testing

gtx-1060-nv-reg-testing

Deus Ex was reported among the Linux games exhibiting slower performance with the 38x.xx driver releases. But at least with the GTX 1070 that sports 8GB of GDDR5 video memory, the performance was actually faster on this newer driver release. But thanks to the Phoronix Test Suite we can accurately plot a side-by-side video memory consumption comparison between these drivers:
gtx-1060-nv-reg-testing

With the first 1080p run of Deus Ex: Mankind Divided on an affected driver release, 1GB+ of video memory is being consumed compared to running the same game in the same clean boot environment on the 378 driver release.
gtx-1060-nv-reg-testing

While no differences in the GPU utilization... But:
gtx-1060-nv-reg-testing

gtx-1060-nv-reg-testing

And in the other Deus Ex: Mankind Divided runs, clearly higher video memory use.

With the GTX 1070 it was also clear in many other Linux games tested that video memory use is going much higher on the current NVIDIA driver:
gtx-1060-nv-reg-testing

gtx-1060-nv-reg-testing

gtx-1060-nv-reg-testing

gtx-1060-nv-reg-testing

gtx-1060-nv-reg-testing

Clearly a problem at play, but with the GTX 1070, the overall performance was fine -- likely due to the 8GB of GDDR5 vRAM and so not coming under video memory pressure with the games tested. Presumably with the lower tier cards it will be easy showing performance regressions if this elevated vRAM use is indeed everywhere. I am still testing other graphics cards, but given the range of tests being used (the ones on this page are just a sampling), it's taking a fair amount of time. I should have more data completed during the weekend.
Related News
About The Author
Michael Larabel

Michael Larabel is the principal author of Phoronix.com and founded the site in 2004 with a focus on enriching the Linux hardware experience. Michael has written more than 20,000 articles covering the state of Linux hardware support, Linux performance, graphics drivers, and other topics. Michael is also the lead developer of the Phoronix Test Suite, Phoromatic, and OpenBenchmarking.org automated benchmarking software. He can be followed via Twitter, LinkedIn, or contacted via MichaelLarabel.com.

Popular News This Week