2 ms·
Something's seriously wrong with your system. I just launched nvim on mine and couldn't measurably detect the interval between hitting enter to exec the command
by kstrauser 2y ago
Something's seriously wrong with your system. I just launched nvim on mine and couldn't measurably detect the interval between hitting enter to exec the command and seeing the welcome screen. Even opening and immediately quitting my substantially customized Emacs took 1.7s.
Your setup should not be taking many seconds to do those routine things. Last time I experienced something like that, I had a dying harddrive that was logging a continual series of read timeouts.
- pdimitar 2y agoThat it shouldn't be like that we all agree with. My Neovim is heavily customized (AstroNvim) but again, it starts instantly on a supposedly weaker CPU. Any pointers on how do I find those read timeout events or any signs of dying hardware?
- kstrauser 2y agoI don’t have that one anymore to test it. I’d suggest using Disk Utility to check the entire drive for errors and ruling that out first. I don’t want to sound like “it works on my machine”, but what you’re describing is so far out of expectations that I’ve gotta wonder if it’s an error condition.
- pdimitar 2y agoI think it's likely the impact of the mitigations against Spectre and Meltdown (iMac Pro uses older-generation Xeon W CPUs). And yeah I'll do the usual checks, disk health included. Been putting it off for a while anyway.
- kstrauser 2y agoI really don't think it'd be those mitigations. Estimates I've seen floating around estimate it to be a total 15-25% performance hit when heavily loaded. For additional context, last weekend I went around to every old Intel Mac in a medical office to upgrade their OSes and some of the apps on them. None of them were speed demons, but they were all just proportionally slower than my much newer Mx Macs. Regular "small" apps still loaded quickly and were perfectly usable. This is in a busy office where any slowdowns that kept people from working at full speed would be fixed. And just because it works for me doesn't automatically mean it's got to work for you, of course. I'm not going to doubt your experience. It's more that what you're describing is so very different than what I'm seeing that it feels like there's got to be something more at play here.