No, I wouldn't use Qlab remote as an operators remote for performance. Although for certain types of application it's not necessarily a complete no no.
To put this into perspective, I have just opened a single projector video show with sound cues in Qlab3
I have 2 iPads connected on my office wifi network. 1 running Qlab remote and 1 running screens 3 in with just the operator screen selected.
Regardless of which machine I trigger the Gos on there is no visible performance hit on the main Qlab machine.
The iPad with Screens 3 is pretty much updating in real time. The hundredths of a second is going like the clappers in the action column and there is no significant delay. This is of course an alternative to Qlab remote if you really need to see the timers (editing in the inspector is nigh on impossible though on an iPad screen with fat fingers.)
Now streaming netflix to my phone on the same network as well. No visible performance hit.
And now to screw it up I change the screens 3 program to display the operators screen and a 2560x1440 monitor with the Qlab video output on it.
Triggering GO is still near instant. There are the odd dropped frames on the video and the iPad with screens 3 is about 1 sec behind and has dropped its frame update to 2fps.
But with a complex sound only show there is no discernible performance hit on the main system regardless of what I try.
I can even put quicktime player on my second screen playing a 1080p video file. Qlab is still fully responsive on the main machine and all remotes and there are no sound glitches.
On the iPad running Screens3 the screen update on the iPad for quicktime player is slow and delayed but the remote screen for Qlab is still updating pretty much instantly