Benchmark Result Discussion

[AshArm]

Desktop and VR Benchmarks on a fresh install and no performance patch, no overclocking. 5080 GPU, 7 9800X3D CPU. Quest 3 with Virtual Desktop used for VR with SSW disabled. I also disabled Asynchronous Spacewarp in the oculas app.

Could you show the performance of the 9800X3D and 5080 at 4K? Thanks.

 

[mmbr20]

Hello. I wanted to share some of my anecdotal findings that you should take with a grain of salt.

I was on a 14900KS paired with a 4090. I also overclocked my ram to 8200 c38. With this setup, I found HT disabled with ecores enabled performed best along with only pcores with HT enabled. HT enabled+ecores I observed a small performance hit.

I am now on a 9800X3D with per core PBO tuning paired with a 5090. I have also tuned my ram here, 2x48gb sticks at 6200 c28. On this setup, I noticed that I am able to push my resolution and MSAA further, but on heavy physics scenes, I am still always limited by the CPU.

I enjoy heavy LD scenes and I'll experience drops to 40 fps using the Quest 3. I assume its because something is being limited to a single thread similar to a game like Microsoft Flightsim. Am I safe in this assumption or could I be doing something more to improve my fps?

(I would post a benchmark result, but I can't get it complete)

 

[NameUsername]

That is what i expected. A 5090 won't do much more than a 4090 in VR with a bunch of things set up (normal amount of assets/plugins/etc.). Thanks for finally letting us all know who were wondering.

 

[Seraphim]

Hello. I wanted to share some of my anecdotal findings that you should take with a grain of salt.

I was on a 14900KS paired with a 4090. I also overclocked my ram to 8200 c38. With this setup, I found HT disabled with ecores enabled performed best along with only pcores with HT enabled. HT enabled+ecores I observed a small performance hit.

I am now on a 9800X3D with per core PBO tuning paired with a 5090. I have also tuned my ram here, 2x48gb sticks at 6200 c28. On this setup, I noticed that I am able to push my resolution and MSAA further, but on heavy physics scenes, I am still always limited by the CPU.

I enjoy heavy LD scenes and I'll experience drops to 40 fps using the Quest 3. I assume its because something is being limited to a single thread similar to a game like Microsoft Flightsim. Am I safe in this assumption or could I be doing something more to improve my fps?

(I would post a benchmark result, but I can't get it complete)

Use the cpu patch to bypass that bottleneck, and you 5090 will perform Much better. https://hub.virtamate.com/resources...to-30-faster-physics-up-to-60-more-fps.43427/

 

[Seraphim]

I just finish building my new PC, I guess it supposed to be a beast but... it seems like I'm missing something because 4K and 1080p are not too far away from each other.
No overclocking, no change of any settings on 3rd party software or hardware beside XMP enabled on Bios, everything is so new that I barely have the machine on for a day now.

I guess this could affect some "extra points" but I'm not 100% sure:
The uptime of the PC is already 4+ hours (I installed windows, drivers, softwares etc..) before I tried VAM's benchmark so it wasn't "chill" maybe without using the PC before I could earn some extra points, but oh well...


4K res:
View attachment 469693

1080p res:
View attachment 469721

See Above. But 285k gonna be tough on you.

 

[T_bob]

Could you show the performance of the 9800X3D and 5080 at 4K? Thanks.

Sorry not possible.

 

[henshin]

Is possible to make some summary of all these data? maybe some table (sadly because there are images is difficult to make with AI)
From what I see 4080 is about 220FPS and 5090 is like 250FPS?

 

[Yankees1550]

Anyone have a 9070xt to test yet?

 

[Case]

Anyone have a 9070xt to test yet?

Soon^TM

 

[ningjiagame]

Waited so long for a 9070XT result, since no one ever posted it, finally bought one myself 囧 (still, with 25% premium over MSRP...), upgrading from 3090, as an interim for future MSRP 5090.
Let's say, the REAL PERFORMANCE of VAM with 9070XT is definately AMAZING, about 33%-50% more FPS comparing to 3090, align with 3DMark scores, and other games like FF7Re.

But very strangely, despite the real performance boost, the CPU benchmark result are almost the same, sometimes even worse than my old 3090:
(9800x3d, 6000/c30, CPU Patch APPLIED)

【1080p】
9070XT:

3090:

【4K】
9070XT：

3090:

【VR】
9070XT only：

Don't know what could have caused this.

Besides, AFMF seems not working, it says "afmf not compatible with the game's display mode".
Other hand, according to some Chinese VAM forum posts, some claims that Nvidia's smooth motion works for vam, even for VR (with cable link only, not virtual desktop streaming).
Anyone with 50series has tested Nvidia's smooth motion with VR? If it really works, it may finally be "the VR version of lossless scaling", and maybe a game changer.

First day, haven't tested everything that carefully yet, may dig further later.
FYI.

 

[Calvados]

Waited so long for a 9070XT result, since no one ever posted it, finally bought one myself 囧 (still, with 25% premium over MSRP...), upgrading from 3090, as an interim for future MSRP 5090.
Let's say, the REAL PERFORMANCE of VAM with 9070XT is definately AMAZING, about 33%-50% more FPS comparing to 3090, align with 3DMark scores, and other games like FF7Re.

But very strangely, despite the real performance boost, the CPU benchmark result are almost the same, sometimes even worse than my old 3090:
(9800x3d, 6000/c30, CPU Patch APPLIED)

【1080p】
9070XT:
View attachment 475660
3090:
View attachment 475672


【4K】
9070XT：
View attachment 475661
3090:
View attachment 475673

【VR】
9070XT only：
View attachment 475667

Don't know what could have caused this.

Besides, AFMF seems not working, it says "afmf not compatible with the game's display mode".
Other hand, according to some Chinese VAM forum posts, some claims that Nvidia's smooth motion works for vam, even for VR (with cable link only, not virtual desktop streaming).
Anyone with 50series has tested Nvidia's smooth motion with VR? If it really works, it may finally be "the VR version of lossless scaling", and maybe a game changer.

First day, haven't tested everything that carefully yet, may dig further later.
FYI.

Hi,According to AMD, only the logging function built into AMD Software can measure the correct frame rate, including intermediate frames generated by AFMF.

 

[ningjiagame]

Hi,According to AMD, only the logging function built into AMD Software can measure the correct frame rate, including intermediate frames generated by AFMF.

Thanks, I got that, my issue was that AMD software shows clearly that "afmf not compatible with the game's display mode" and definately not working.

And, I finally found a way to active AFMF 2.1 in VAM, with 9070XT:

---- by forcing VAM to run with 【DirectX11】!

(by adding "-dx11" in VaM (Desktop Mode).bat)

Also, it 【works in VR】!! （maybe only works with cable link PCVR headsets and won't work with WIFI streaming headsets / virtual desktops）
Managed to run threesome sceens at the max refresh rate of 90Hz (45Hz+AFMF) with my HP Reverb G2, at SteamVR resolution set to 100%.
(You may need to set both the physics rate in VAM and Radeon Chill refresh rate to half of max refresh rate of your head set to get correct animation speed)

Further, in most heavy sceens, you can even further turn on SteamVR motion smoothing, it works together with AFMF, so you can get 3~4 times more smooth animations (though obviously with poorer quality) , and have definately different VR experience!

Since my old 3090 only got 10~fps in most heavy sceens... I will definately say, the VR applicable AFMF is a game changer for me ---- again, finally "the VR version of lossless scaling".
(Don't know whether it works on 6000 series or earlier AMD cards. Nvidia 50series' smooth motion is also reported to work in VR, but I don't have one so cannot testify.)

FYI

 

[VAMSOY]

Here is a direct comparison between 4090 and 5090 @4K, everything else was kept the same, no OC at all.

Note that this is done on an older CPU (13900K), with a clean install of VAM. The improvement is larger than I expected since VAM doesn't take advantage of DLSS4.

 

[Wxy_332]

Here is a direct comparison between 4090 and 5090 @4K, everything else was kept the same, no OC at all.

Note that this is done on an older CPU (13900K), with a clean install of VAM. The improvement is larger than I expected since VAM doesn't take advantage of DLSS4.

View attachment 476768View attachment 476769

that's really useful, almost double the 4k performance.
How about the VR expericence, does it double also?

 

[Case]

Thanks, I got that, my issue was that AMD software shows clearly that "afmf not compatible with the game's display mode" and definately not working.

And, I finally found a way to active AFMF 2.1 in VAM, with 9070XT:

---- by forcing VAM to run with 【DirectX11】!

(by adding "-dx11" in VaM (Desktop Mode).bat)

Also, it 【works in VR】!! （maybe only works with cable link PCVR headsets and won't work with WIFI streaming headsets / virtual desktops）
Managed to run threesome sceens at the max refresh rate of 90Hz (45Hz+AFMF) with my HP Reverb G2, at SteamVR resolution set to 100%.
(You may need to set both the physics rate in VAM and Radeon Chill refresh rate to half of max refresh rate of your head set to get correct animation speed)

Further, in most heavy sceens, you can even further turn on SteamVR motion smoothing, it works together with AFMF, so you can get 3~4 times more smooth animations (though obviously with poorer quality) , and have definately different VR experience!

Since my old 3090 only got 10~fps in most heavy sceens... I will definately say, the VR applicable AFMF is a game changer for me ---- again, finally "the VR version of lossless scaling".
(Don't know whether it works on 6000 series or earlier AMD cards. Nvidia 50series' smooth motion is also reported to work in VR, but I don't have one so cannot testify.)

FYI

Hmmmmmh ... that's really interesting, thanks for pointing this out. Maybe time for a tutorial post? I don't think I've seen this mentioned anywhere else on the Hub.

 

[Seraphim]

that's really useful, almost double the 4k performance.
How about the VR expericence, does it double also?

Thats +62%. But yes typical in vam vr is 40-60%. Getting +53% in this bench.

 

[donthavethetime]

Got a 5090 so here are my results. I moved from a 4080 super previously which was good, but left some to be desired and wasn't allowing me to run Voxta well with Vam. I've been able to max out everything, but FPS in scenes so far hasn't been mind blowing (needs more testing). Curious what others are seeing as my 5090 barely runs at 40% most of the time.

I tried a few things for each of them I did the OSW disable, turned off in Steam, and turned off VD ASW. confirmed i was seeing 130-150fps in a scene.

ALERT: Just an FYI as I was an idiot and you might be too. I ran 3dMark and got a depressing score originally. Looked online and someone mentioned doing a clean install of drivers even though I had latest. Figured wth and saw a 25% increase in score and was over average without any overclocking or anything. So definitely do this. was in the high 11k range and went up to almost 15k. 3dmark demo.

First, VirtualDesktop ran directly in VD no steam. not sure why it says rift, but on Quest 3.

Then I tried my oculus script -vrmode oculus which launched steam.

Then I ran the default -vrmode openvr which ran steam.

Also ran a desktop at 1440p.

 

[dogesnvd]

motherboard MSI Z790-P D5,
M.2 WD BLACK SN770 2T,
DRAM 32GB DDR5 6400 C32*2
13600KF+4090, 4k desktop NO CPU Performance Patch .and no OC,

13600KF+4090 4k desktop with CPU Performance Patch No cores or hyperthreading turned off.and no OC

so min1% is a little improve.This is more noticeable in high-stress environments and VR.Movement is smoother.

13600KF+5090D/5090 (Theoretically there is no difference in performance） desktop with CPU Performance Patch No cores or hyperthreading turned off.and no OC ,Same test environment as RTX4090

AVG FPS improved 56.8%, min1% improved 60.1%, max1% improved 80.7%.

13600KF+4090 VR with CPU Performance Patch No cores or hyperthreading turned off.and no OC

13600KF+5090D/5090 (Theoretically there is no difference in performance） VR with CPU Performance Patch No cores or hyperthreading turned off.and no OC ,Same test environment as RTX4090

AVG FPS improved 58.6%, min1% improved 49.3%, max1% improved 46.6%.

If you have the same doubts as me about the improvement of 5090/5090D.
You can refer to my results. I think upgrading to 9800X3D or 9950X3D will get better frame rate performance and min1% low.
For complex physical calculations, such as the two-person MMD dance scene, 13600KF is not enough to support sufficient physical calculation frames.