Tested: Default Windows VBS settings slow down games by up to 10%, even on RTX 4090

Remember when Windows 11 was released and there was a concern about how enabling virtualization-based security (VBS) and hypervisor-enforced code integrity (HVCI) by default might affect performance? A lot of noise was made, benchmarks were run…and then we all just moved on. Fast forward to 2023, and I recently discovered that sometime in the last few months, the PC I use for the GPU benchmarking hierarchy received an update that turned VBS back on. (We have an article on how to disable VBS if you want.)

Windows 10 also has this setting and it may now be enabled by default as well. Tom’s hardware editor-in-chief Avram Piltch uses Windows 10 Home on his main desktop and found that VBS seemed to be enabled even though he never touched the settings and had clean Windows installed over the summer.

This default VBS settings everywhere worried me, because I’m already in the middle of retesting all relevant graphics cards for the 2023 version of the GPU hierarchy, on a new benchmark that includes a Core i9-13900K CPU, 32 GB of DDR5-6600 G.Skill Memory and a Sabrent Rocket 4 Plus-G 4TB M.2 SSD. Needless to say, it doesn’t put together best-in-class parts just to run extra features that can affect performance.

Except… that I did. When I prepared the new benchmark for next year in November, just before the release of the RTX 4080 and RX 7900 XTX/XT, I was short on time. I installed and upgraded Windows 11, downloaded the rather massive 1.5TB of games I use for testing to the SSD, and got to work, all with VBS enabled. Having caught my breath and given a bit of extra time, I belatedly realized my mistake, if you can call it that.

So I set about testing and retesting the performance of the fastest graphics card, the GeForce RTX 4090, with and without VBS enabled. After all, we’re now two new generations of CPUs beyond what we had at the launch of Windows 11, and with faster CPUs and new architectures, perhaps VBS will have even less of an impact than before. At the same time, we are also using new GPUs that provide substantially higher performance than the RTX 3090, which was the fastest GPU in 2021, which could lead to CPU bottlenecks and extras like VBS. further of an obstacle than before.

Windows 11 VBS Test Hardware

You can check out our test PC hardware, using Nvidia’s 528.49 drivers (which have now been superseded three times). Let’s get straight to the results, with our updated test suite and setups consisting of a battery of 15 games, across four different settings/resolution combinations. Let’s summarize things in a table, divided into Average FPS on the left and Low 1% FPS (the average FPS of the bottom 1% of frames) on the right.

To be clear, all tests were done on the same PC, over a period of a few days. No game updates applied, no new drivers installed, etc. to keep things as apples to apples as possible. The only change was to disable VBS (because it was on initially, the Windows 11 default).

Each test was run multiple times to ensure consistency of results, which raises the only discrepancy: Total War: Warhammer 3 performance is all over the place right now. I don’t recall that having been the case in the past, but sometime in February or perhaps early March, things seem to have taken a turn for the worse. I’m still investigating the cause and I’m not sure if it’s the game, my system or something else.

(Image credit: Tom’s Hardware)

Taking the high level view of things, maybe it doesn’t look too bad. Disabling VBS improved performance by up to 5% overall, and that dropped to just 2% in 4K ultra. And if you’re running this level of gaming hardware, we think you’re probably expecting to run 4K ultra as well. But even at our highest possible settings, there are still some notable exceptions.

The biggest overall improvement comes in Microsoft flight simulator, which makes sense since that game tends to be very CPU limited, even with the fastest possible processors. Disabling VBS consistently improved performance in our RTX 4090 tests by around 10%, with the lows of 1% increasing by as much as 15%.

Not by chance, flight simulator it’s also one of the games to absolutely love AMD’s great 3D V-Cache on the Ryzen 9 7950X3D. Our CPU benchmarks use a different, less demanding test sequence, but even there, AMD chips with large caches are 20% (Ryzen 7 5800X3D) to 40% (7900X3D) faster than the Core i9-13900K. Perhaps VSB would have less of an impact on AMD’s X3D CPUs, but I didn’t have access to one of them for testing.

Another game that tends to run into CPU bottlenecks at lower settings is far away 6, and it also saw fairly consistent 5% or more increases in performance, noticeable in benchmarks but less so in actual gaming. Curiously, cyberpunk 2077 with ray tracing enabled, 5% higher performance was also observed. Perhaps this is because the work of building the BVH structures for the ray tracing computations happens on the CPU; many of the other ray tracing games also showed increases of 5% or more.

Microsoft flight simulator

(Image credit: Microsoft)

What about games where VBS didn’t matter much, if at all? infinite bright memory (the standalone benchmark, not the full game) showed almost no changes, and Minecraft It only showed a modest improvement at 1080p with our most demanding settings (24 RT render chunk distance). A Plague Tale: Requiem, borderlands 3, force horizon 5and red dead redemption 2 it also showed less impact, although in some cases the minimum FPS may have changed more.

(And again, I’m not really saying anything about Total War: Warhammer 3 since the performance fluctuated too much. Even after 20+ runs each, with and without VBS, there was no clear typical result. Instead of a bell curve, the results fell into three groups in the low, medium, and high range, with the lows of 1% showing even less consistency. removing TWW3 However, of our geometric mean, only the low 1% delta changes by less than two percent, so I left it.)

The larger deltas are generally at 1080p, and it didn’t seem to matter much if we were running “medium” or “ultra” settings. This is probably because ultra settings are often more CPU-hungry for other computations, so it’s not just higher resolution textures or shadows.

But the question remains: to VBS or not to VBS? Especially for my GPU tests. The good news is that it’s pretty much a never-ending process, as new drivers and game patches seem to routinely invalidate previous results. It might at some point switch to having VBS disabled, and maybe it will. But that new test is also the bane of GPU benchmarks.

Windows VBS: The Bottom Line

Source link

James D. Brown
James D. Brown
Articles: 8277