Prateek Shukla

I updated the script as you said the main role of PR was to eliminate the crash in SDDM which is resolved

> I do not approve of this PR at all, for the following reasons: > > 1. The previous behaviour of `bar.sh` was to check for updates during the first...

> pure posix! I appreciate your emphasis on speed and everything but the problem is that it crashes sddm on XF86XK_AudioMute. I don't even know why I tried everything in...

> Anyway, OP should be more specific so we can make sure dash is the problem indeed (which I doubt). Initially I doubted it too, so I started removing each...

I am getting error but it is not conclusive - Dec 27 23:14:23 archlinux sddm[841]: Session started true Dec 27 23:14:23 archlinux sddm-helper[462855]: [PAM] Closing session Dec 27 23:14:23 archlinux...

> I don't see any error, it seems like chadwm stopped, one way or another. Did you try the steps I mentioned? Yeah, I tried what you said, I reinstalled...

https://drive.google.com/file/d/1LNsLcBMX-lLiIc7myJm8AtRuXZFzcIzV/view?usp=sharing here's the video. I also included the same activity in another window manager to show that this is not because of some driver issue or something else but issue...

I think its because a (1,1) thread layout local_partition is only giving each thread access to one element per dimension of the tensor rather than the full tensor data you...

When you call local_partition it divides your tensor according to the thread layout you provide. When your thread layout was configured as (1,1) or improperly sized local_partition would give each...