Benchmark Result Discussion

[Jiraiya]

To mirror what others have said, 16GB is probably absolutely fine for most gamers. 32GB is nice, but in no way needed for VaM.
You will almost certainly have a pair of 2 * 8GB modules if you have 16GB currently. This means a fairly easy upgrade to 32GB if needed (add 2 more of the exact same memory you have) but will not gain you much if anything in VaM.

 

[Jiraiya]

Here's what I got. Not much but the 1070 is still kicking and doing it's best! It's a real trooper. <3

View attachment 72357
I'm running Windows 11 [ version: 10.0.22000 Build 22000 ] and I have some overclock in place:

CPU: Running at 4625MHz 1.325V

GPU: +128MHz on the core and +300MHz for the memory.

I have the same CPU as you and a 3060. I am getting <1fps more than you. It's close enough I wouldn't say it counted.
Your 1070 is really pulling it's weight considering the age The only thing missing is the RTX part, but honestly nothing uses that at the moment (certainly not VaM) so no real loss.

 

[fabi88]

My result (Desktop Mode):

Specs:
CPU Brand Name: Intel(R) Core(TM) i7-4790K CPU @ 4.00GHz
Motherboard Model: ASRock Z97 Extreme6/3.1
Motherboard Chipset: Intel Z97
Video Card: ZOTAC GeForce RTX 2080 Ti AMP/Triple Fan
Video Memory: 11264 MBytes of GDDR6 SDRAM [Micron]
Monitor Name (Manuf): LG ULTRAWIDE
Drive Model: Samsung SSD 970 PRO 512GB
Memory Module Manufacturer: Corsair
Total Memory Size: 32 GBytes
Memory Type: DDR3 SDRAM
Memory Speed: 1200.5 MHz (DDR3-2400 / PC3-19200)
CPU-Temp on Test: ~ 50° C
GPU-Temp onTest: ~ 65° C (MSI Afterburner in Background)

 

[Bob Nothing]

Not too sure. Could be the NVMe drive?

I run on NVME as well

 

[HolySchmidt]

NVMe has nothing to do with this benchmark. All data needed is loaded in the GPU VRAM and in normal RAM. Right after you click on load scene. If someone has problems with unsteady stuttering THEN you have too little VRAM or RAM. NVMe only has an impact on how fast the scene is loaded.

 

[Bob Nothing]

NVMe has nothing to do with this benchmark. All data needed is loaded in the GPU VRAM and in normal RAM. Right after you click on load scene. If someone has problems with unsteady stuttering THEN you have too little VRAM or RAM. NVMe only has an impact on how fast the scene is loaded.

Agreed. I have 32GB 3200mhz RAM and 24GB of VRAM, so no issues there

 

[geo_gan]

To mirror what others have said, 16GB is probably absolutely fine for most gamers. 32GB is nice, but in no way needed for VaM.

Well actually you are slightly off there it is more than nice... VAM is an absolute memory hog... for some reason it does not clear its memory when changing scenes... and will usually just crash because of memory problems very soon after a few scene changes. I often see VAM using up way more than 16GB memory if its there... some stress scenes went to 32GB on mine before crashing. VAM is way more memory dependent than most games which are specially designed to fit in some standard memory amount like 8GB.

 

[TToby]

@geo_gan
you are absolutely right, but most users don't have that crashing issue, or at least not that often.
Maybe they are not using that much scenes/textures per VaM session. I did well with 16GB in VaM for several years and only 1-2 crashings per week..
I have buyed my own additional 16GB bar only some month ago out of a stupid "will have" mood. No differences, but also no crashings.
I would say: If you don't have issues with VaM crashing more or less often, then "you" (the undecided reader) are fine with 16GB.
For whom it may concern: there is a button in the figures main menu front tab called something like "delete unused textures". If you will press this button once in a while after some havy clothes browsing or scene loading, you are fine, too.

 

[Jiraiya]

Oh, I certainly get the odd crash. I just reload and carry on.
16GB is enough to run VaM, load a scene, use it. If you keep reloading new scenes without closing you will get a ram related crash eventually. I don't think that is indication you need more ram, but that VaM needs some fixes....

 

[Ben-g Copper]

Thank you for developing this testing tool. Compared to other people's test results, it gave me a better understanding of what hardware upgrades should be.

 

[HolySchmidt]

Seeing more and more people throwing their benchmarks in, why are you people not changing at least your desktop resolution to a normal scale like 1920x1080?? I mean we are here for comparison! Everyone got their own new desktop resolution scale, this makes comparing a little confusing and difficult!

1845x1057
1920x1009
1920x1017
2560x1057
3840x1600

I mean it's a bit of a mess.

 

[VeeRifter]

Here's my benchmark results:
VR - Resolution Scale 1x
Overall FPS 68.54

VR -Resolution Scale 2x
Overall FPS 52.33

Desktop 1920x1080
Overall FPS 157.56

Desktop 3840x2160
Overall FPS 76.03

 

[Kawum]

Hi. Was a bit of a pain getting the benchmark to work. Ended upwith a clean install as my std install needs some serious decluttering.
1080p for some reason is a no go on my system.
OC is 5 on all cores, mem 3466, GPU custom curve

 

[popov]

I have a Quest 2 and use Virtual Desktop streamer but I can only get 20-21 fps for both avg, min and max framerates.
This is constant during all 5 tests, seemingly independent from how much the scene demands.
I disabled asw with the .bat file in the attached zip to this plugin.
Not sure if this is what to expect with my old system, or if something is very wrong in my settings?

 

[MacGruber]

Not sure if this is what to expect with my old system, or if something is very wrong in my settings?

VaM has a lower limit of 20fps, if your PC can't make that the game will slow down instead, messing up the benchmark results. I have some ideas how to fix that in the next version of the benchmark. It will contain a 3 character scene then, so more people will hit the 20fps in VR, so it becomes more relevant to fix it.

 

[mediamac1]

Many thanks @MacGruber for this superb peice of code. VaM Benchmark is super useful and I'll try to optimize some of my system settings when running under Quest2/VRDesktop next few days using the results of the benchmark.

 

[Jiraiya]

I think what this is telling me is that everybody has a better PC than me and you all managed to get 30xx cards a damn site easier than I did!

 

[HolySchmidt]

I think what this is telling me is that everybody has a better PC than me and you all managed to get 30xx cards a damn site easier than I did!

I can imagine how you feel. Driving around with my car I sometimes have the feeling all people have more expensive, newer and better cars than me. But at the end I'm lucky to at least have a car. What do you think how many people on this earth don't even own a computer? So, always try to be happy with what you got, not be unhappy about what you could have.

@VeeRifter
Thanks for sharing. Owning a 9700K with 3080 (all stock) I got ~140fps at 1920x1080. Did you overclock something? Or is this the difference between 9700K and 9900K? You have Hyperthreading on? I mean that's more than 10% better ... should I?

btw: No one in here owning a 12x from Intel? That would be really interesting right now. 12700K/12900K + 3080/3080ti/3090.

 

[MacGruber]

Version 3 has been released!
There have been changes to lighting, more props and so on as well as 2 new scenes were added.
Results from earlier versions are not comparable with version 3!

Changelog: https://hub.virtamate.com/resources/benchmark.11336/updates#resource-update-15286

 

[MacGruber]

New results for benchmark version 3.

My system is using DDR4-3200 CL14 memory, which should in theory be roughly 16/14 = 14% faster than the usual cheaper CL16 memory. Also the CPU is undervolted which gave me about 10°C less with some 1-2% more performance in 3DMark. The graphics card is an MSI GeForce RTX 3070 GAMING Z, so slightly overclocked like most third-party cards.

It's noteworthy that apparently resolution scale in VR does not make a difference on my machine. Too lazy to do another VR run with 1.0x resolution scale, but it's almost identical to running with 1.5x for me. The mirror scene as all well as both new 3 character scenes seem to lock to 45fps in VR, probably because Oculus forces VSync, so that sub-result may not be useful.

 

[gokusanmar]

You guys can get over 100 frames in this!?

That's why I stick to only two models and absolutely nothing in the background.

 

[MacGruber]

@gokusanmar Interesting result for Baseline3 / SimplerPhysics. The FPS did not improve much for some reason, but the time measurements did improve a LOT.

 

[gokusanmar]

Since I am using a laptop, I am interested in checking out the difference between running VAM on a laptop screen vs monitor ( supposedly games run ~10% better when you have a monitor plugged if your laptop does not have MUX switch) and then 1080p vs 1440p.