Ok, I've done some very extensive testing, benchmarking everything before and after the lag starts with various different settings. I've learned the following. GTX 580's throttle in FurMark, making the score much lower than expected. However, this throttling happens less often while the VM is lagging, causing the scores to be significantly higher during the lag. RAM, CPU, and SDD R/W speeds all drop slightly during the lag regardless of the other settings. Dedicating processor cores to the VM makes little difference, although it performed slightly better on average. CPU cores had to manually assigned to the vcpu child processes with taskset, when I tried to assign the range 0-3 to the master process it only used the first core and obviously lagged badly until I manually assigned the other cores. The problem is consistently triggered by interacting with the Qemu Monitor window in BSPWM. It's been triggered by switching to or away from its workspace, mousing over it, and moving it in any way. When I interact with it its font grows significantly larger. Upon switching to -nographic and changing nothing else I don't seem to be having any more stuttering issues at all. I haven't extensively tested yet but previously I could get the issue to trigger 100% of the time by moving the monitor window. Moving the nographic terminal window doesn't have this issue. Does anyone know why this could be happening? It seems like a really unusual cause for this kind of problem but it really makes sense when I think about the situations that usually triggered the stuttering (Opening and moving windows, moving my mouse across the screen while using synergy) and that it never happened when I was running with my graphics drivers disabled. Maybe I'm onto a bug here, I have definitely not seen anyone else with this issue in my research. Someone please let me know if this is a know problem, I'll use the VM for a while and update if I have any other performance issues. <div class="gmail_quote">On Feb 28, 2016 12:25 PM, "Ben J" <<a href="mailto:btpprograms@gmail.com">btpprograms@gmail.com</a>> wrote: <blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">It seems that CPU usage isn't the issue from further testing. I used isolcpus to disable cores 0-3, launched the VM with only those cores in use and each dedicated to one vcpu (used taskset and manually bound each core to one of the four child processes that had cpu usage), and was still able to get the same issue to occur. Once the VM slows down it doesn't ever speed up again without a complete shut down. That's the main thing that confuses me, it's one thing for the VM to slow down from high resource usage but it continues lagging even when it's just sitting on the desktop. It seems to get triggered pretty often by workspace switching and window movement in BSPWM but that's really the only lead I've got at this point. The issue still seemed to happen even when I disabled audio altogether. Anyone have any other ideas? <div class="gmail_quote">On Feb 27, 2016 12:36 AM, "Ben J" <<a href="mailto:btpprograms@gmail.com" target="_blank">btpprograms@gmail.com</a>> wrote: <blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr">I've recently been having issues with my VM's framerate dropping at random. It generally boots up completely fine, but at seemingly random points it starts to stutter and doesn't stop until the VM is fully turned off and back on again. It started out being fairly rare, but now happens on almost every launch of the VM after a few minutes. My specs:Host OS: Arch LinuxGuest OS: Windows 7Motherboard: Asus Rampage IV Extreme CPU: I7-3960X Graphics: 2X EVGA GTX 580'sMy Launch Script:killall synergyssynergys QEMU_PA_SAMPLES=4096 QEMU_AUDIO_DRV=pa qemu-system-x86_64 -enable-kvm -m 4096 -mem-path /dev/hugepages -cpu host,kvm=off \-smp 6,sockets=1,cores=6,threads=1 \-bios /usr/share/qemu/bios.bin -vga none \-device vfio-pci,host=02:00.0,multifunction=on,x-vga=on \-device vfio-pci,host=02:00.1 \-net nic,model=virtio,vlan=0,macaddr=DE:AD:BE:EF:D4:41 \-net tap,vlan=0,id=network0,ifname=tap0,script=no,downscript=no,vhost=on \-usb -usbdevice host:046d:c534 -usbdevice host:045e:028e \-soundhw hda \-boot menu=on \-drive file=/dev/sda1,format=raw,if=virtio,cache=none \-drive file=/mnt/LinuxStorage/ISOs/Windows/Drivers/virtio.iso,id=virtiocd,if=none,format=raw -device ide-cd,bus=ide.1,drive=virtiocd The issue seems to get triggered immediately whenever I launch a Skype or Discord call. It also seems to be more likely when I'm running a game within the VM (Tested with Rocket League and Rivals of Aether). It has also happened without either of those being open though. I'm really at a loss. I've tried disabling huge pages since I suspected that I might not have enough, but that showed no difference. Restarting the computer fixes it temporarily but it usually happens again shortly after I restart. The same thing happens when I suspend and wake up the computer. Has anyone experienced this issue before? </div> </blockquote></div> </blockquote></div>