Hello !
I am still having a performance regression between the a11 and A12 version of Looking Glass
My specs :
9700K, 6 cores to the guest and 2 for the host (I tried 4/4, it did not change anything)
2080 TI
1920*1080 @ 60 FPS
Asrock Z390 Phantom Gaming ITX/AC
Archlinux as the host OS. I’m not using the AUR package for the client, but compiled it from the sources on github.
Native (Screen directly connected to the GPU) : 1046 Cinebench score
A11 : 966 cinebench score, no difference in gaming
A12 : 841, but suddenly all games lose a ton of fps, WoW can’t maintain 60 locked fps anymore and goes down to 40/45, more modern games like Assassin’s Creed Odyssey become unplayable, etc.
CPU usage of the host application is normal : 2.4%.
I get my framerate back as soon as I close the client application on Linux side.
You are one of the very few people using a RTX card Looking Glass, as I do not have one to test with I can not replicate this. I do not see these issues on a 1080Ti, so it is likely an issue with the later generation cards.
I have some issues with FPS aswell using looking glass, I’d like to make some improvements to make it better, hence my question.
Let me start with my pc specs and setup:
I7-8700 (non K)
16gb Ram
GTX 1080 Ti
Running a windows VM on the GTX which runs my game fullscreen at 90 fps+
My linux which I run Looking glass from is connected to my igpu, once I start looking glass, frames drop to about 30 - 40.
The question here is of course ‘why?’.
But more important, how can I improve this, is it lack of ram, cpu power or simply because LG runs on the igpu and not on a second GPU?
Because once you start the client the host starts capturing and sending frames to it.
Windows capture API and NVidia not playing ball with non NvFBC users (aka, not Quadro users).
Are you running your game in full screen windowed mode? or full screen? For an unknown reason running in “Full Screen” mode performs terrible on some titles.
when i type make, it says:
make: *** No targets specified and no makefile found. Stop.
running fedora 29, i installed fedora because i heard that qemu 3.0 included a patch for amd smt bug that allowed amd processors with smt to use it in a vm, im running amd threadripper 1950x and gtx 1080 ti. i saw a 30 fps gain coming from ubuntu which is running qemu 2.16, and yes ive tried to install qemu 3.0.0(on ubuntu), but like using Looking glass on fedora there are very few articles i could find to install it, and when i did install it it just made everything break. So im sticking with fedora, but id like to use looking glass.
after some work i got it installed “[100%] Built target looking-glass-client”
I do however want to note some of the issues i had when installing.
i made myself a crappy bash script to install it and it looks like this
you can run this or
put it in a bash script like i did
next
cd /home/user
git clone https://github.com/gnif/LookingGlass.git
cd ./LookingGlass/client
cmake ./
make
change “user” to your user.
gnif i noticed that in your tutorial you said cmake ../
im not sure why but in fedora i just got this error
“CMake Error: The source directory “/LookingGlass” does not appear to contain CMakeLists.txt.
Specify --help for usage, or press the help button on the CMake GUI.
make: *** No targets specified and no makefile found. Stop.”
but using cmake ./ works fine and thats what i used
touch doesn’t create a directory, it creates an empty file.
You need to do this before starting QEMU, not the client. Qemu’s ivshemem virtual device needs access to this file and will create it. The touch and chmod are simply to set it’s initial permissions so that you can access it.