Virtu MVP, How it works with your GPU..

[rockstar_7]

Hey guys,

I finally setup Virtu MVP for my Gigabyte z77x-d3h, 3570K-660Ti rig (windows 7 x64). It wasn't difficult but it was definitely quite confusing.

Here's how it is setup:
Step 1) BIOS>Peripherals>Integrated Graphics>ENABLE iGPU(instead of AUTO) -- you need to do this to install the Intel HD4000 Driver.
Step 2) BIOS>Peripherals>Graphics Input interface (or something)> iGPU (instead of AUTO or PCI) -- This is required so that you may boot using the Intel hd4000 to install the driver
Step 3) Download and Install Intel HD4000 Driver: http://www.intel.com/p/en_US/support/detect/graphics
Step 4) Download and Install Lucid Virtu MVP setup Driver: http://www.lucidlogix.com/driverdownloads-mvp.shtml
Step 5) Reboot. Your PC will be running using the iGPU now, so be aware of that at all times.

Although i managed to install the Virtu MVP driver and Intel HD 4000 Driver.. i just did not understand how it works... So, heres how i started messing around with Virtu Mvp..

I remember watching a Virtu MVP demo which demonstrated how effective it could be in a gaming environment, and oh boy was I impressed! So, i decided to test my Virtu mvp setup just to see if t was properly configured. The demo used the Super Street Fighter 4 Benchmark and since i wanted to compare my results with that of the demo.. i didnt stop myself from downloading it.

Heres the Demo, just in case:

Spoiler

http://www.youtube.com/watch?v=K7FB1wfM53E

In my case, I ended up playing around with the display mode (switcing between hd4000 and 660ti) and testing the effects of VirtuMVP by running the Super Street Fighter IV Benchmark, only to discover a slight disparity..
I tried both inputs, VGA (Intel HD 4000) and HDMI (660Ti) to test Virtu MVP without having the slightest idea of being able to find any sort of significance in my testing.

So, heres how i had been changing screen modes:

Spoiler

I've got a vga cable (onboard graphics chipset) and an hdmi (660Ti)-- dont have a dvi cable -- connected to the same monitor, allowing me to switch the display mode by right clicking >> 'screen resolution' and selecting (from the drop down box) 'display on screen 1' for HD4000 or 'display on screen 2' for 660Ti with ease.

Here are my observations for the following tested Scenarios:

VGA-MODE INTEL HD4000:

Scenario 1)
GPU Virtualization 'On'
Display mode: Intel HD4000
Ran SF4B: @720p No Vsync
Result: 273fps

Scenario 2)
GPU Virtualization 'Off'
Display mode: Intel HD4000
Ran SF4B: @720p No Vsync
Result: 54fps

As you may have noticed, Virtu MVP is actually doing something rather than just involving the Discrete GPU when being 'On'. It seems like it manages to improve the performance, by activating my discrete GPU (660Ti), quite significantly.

However, it didnt seem to be anything close to satisfactory, and so i decided to change the display mode thinking i'd be better off not using virtu MVP at all and just use my disrete graphics. And Now that i had switched the Display Mode, i just ended up trying it alongside SF4B (again), only to have discovered a much pleasing result.



HDMI-MODE 660Ti:

Scenario 1)
GPU Virtualization 'On'
Display mode: 660Ti
Ran SF4B: @720p No Vsync
Result: 566fps

Scenario 2)
GPU Virtualization 'Off'
Display mode: 660Ti
Ran SF4B: @720p No Vsync
Result: 523fps

I was quite surprised to see the difference to be 2x that of the framerate in VGA Display mode (and more with GPU Virtualization 'On').

Even though i was pleased, i was still quite confused.. since it didnt make any sense -- the behavioral difference, which was directly related to whether the pc was in Intel HD4000 VGA-MODE OR 660Ti...

That's how i came up with this question.... of How Virtu MVP works.. cause this does not seem to be consistent at all..

-- So, Care to Explain?



I even went to the extent of gathering a little more data.. And, This time i tested the Virtual_VSYNC Feature of Virtu MVP:



Vsync

Scenario 1)
GPU Virtualization 'On'
Display mode: Intel HD4000
Ran SF4B: @720p With Vsync
Result: 58fps

Scenario 2)
GPU Virtualization 'Off'
Display mode: Intel HD4000
Ran SF4B: @720p With Vsync
Result: 49fps

Scenario 3)
GPU Virtualization 'On'
Display mode: 660Ti
Ran SF4B: @720p With Vsync
Result: 58fps

Scenario 4)
GPU Virtualization 'Off'
Display mode: 660Ti
Ran SF4B: @720p With Vsync
Result: 59fps

It's all pretty much the same to me, which is pretty sad. Apparently, turning Virtual Vsync does a pretty decent job in order to allow higher framerates.. well, at least the Demo showed that it did.. so i gave that a shot too..


Here's a look at Virtual Vsync and Hyperformace:
(which is done using the 660Ti (HDMI) as its proven that my 660Ti is much more compatible with Viru MVP than VGA-Intel HD4000 Display Mode)

Virtual VSYNC
GPU Virtualization 'On'
Display mode: 660Ti
Ran SF4B: @720p With Vsync
Result: 58fps

Virtual VSYNC & Hyperformance
GPU Virtualization 'On'
Display mode: 660Ti
Ran SF4B: @720p With Vsync
Result: 59fps

Hyperformance
GPU Virtualization 'On'
Display mode: 660Ti
Ran SF4B: @720p no Vsync
Result: 472fps

Once again, Virtu MVP's behavior didnt make sense. With Virtual VSYNC Turned 'ON'.. it didnt make much of a difference. And with 'Hyperformance' activated.. It DECREASED THE FRAMERATE! The Virtual Vsync and Hyperformance results were nothing like that in the demo, they were positive for ****sake.

-- Care to Enlighten me? So How does Virtu MVP work? If it displays such disparities just by simply changing the input mode...

All i understand is, that it might reducing a little load from the discrete Graphics on to the iGPU... which seems to be quite incomplete, yet.

 

[rockstar_7]

Nevermind! I got it work the way it should...

I'm running Windows using my 660Ti HDMI connection instead of VGA HD4000 and I'm using Virtu with Hyperformance and guess what... I GET 680FPS!!

All i had to do was play it in FULL SCREEN! -.-
Anyone know how to take an in-game screenshot in Full screen mode.. cause, thats why i was using a lower resolution and windowed mode in the first place!

If tell me how to.. i'd like to post 2 screenshots:
Hyperformance: ~682 FPS
Virtual Vsync: ~420 FPS -- yes, i get 420 with no tearing.. dont know how.. but, im happy

SO, CONCLUSION.... Virtu MVP works as long as you can get it set up correctly!

 

[Bromeh]

Nice post even if it's a quite old, i just discovered it searching for my VirtuMVP driver and i found in your top links thanks

I benchmarked some games to see if that's going to do a big difference so i will keep you informed on my testings.


- The Brownie

 

[rockstar_7]

Nice post even if it's a quite old, i just discovered it searching for my VirtuMVP driver and i found in your top links thanks

I benchmarked some games to see if that's going to do a big difference so i will keep you informed on my testings.


- The Brownie

I'm glad this post helped you out.
Please do tell me if you discover anything when testing VirtuMVP, that would be nice

 

[Bromeh]

Hi Rockstar, i'm done with my tests and i wasn't surprised at all byt the performances. I used Call of Duty Black Ops 2 in single player during the first mission, then Modern Warfare 3 in multiplayer in the map Dome, Counter Strike Source in a Zombie Mode, Euro Truck Simulator - driving around normally, Far Cry 3 during the first mission, PlanetSide 2 just looking around and shooting and finnally Skyrim during the beggining of the game (on the road to get killed).

My PC specs are :

i7 3770k 4100MHz
GTX 660Ti (Same as you )
8GB 1600MHz
120GB SSD

Here's a little graph i made :

On every game, i was getting lower FPS except on Counter Strike and Skyrim (i guess because they are older). Also, on Skyrim, it wasn't playable. Screen was shacking, everything was moving, i went in the sky and then back down on the road ... I guess it dosen't support it very well.

Anyway, i guess it's not something that the modern gamers want, maybe it works better with older cards, maybe not.


- The Brownie


Edit: Here's the exact framerates of all the games :

Off :

Black Ops 2 - Average:114
Modern Warfare 3 - Average:91
Counter Strike Source - Average:221
Euro Truck Simulator - Average:59
Far Cry 3 - Average:48
PlanetSide 2 - Average:71
Skyrim - Average:57




With VirtuMVP :

Black Ops 2 - Average:96
Modern Warfare 3 - Average:71
Counter Strike Source - Average: 269
Euro Truck Simulator - Average:57
Far Cry 3 - Average:47
PlanetSide 2 - Average:73
Skyrim - Average:63


With VirtuMVP and Vsync :

Black Ops 2 - Average:117
Modern Warfare 3 - Average:74
Counter Strike Source - Average:298
Euro Truck Simulator - Average:56
Far Cry 3 - Average:47
PlanetSide 2 - Average:66
Skyrim - Average:75