Re: Remote development with neovim, tmux and SSH from macOS?
- In reply to: Doug Hardie : "Re: Remote development with neovim, tmux and SSH from macOS?"
- Go to: [ bottom of page ] [ top of archives ] [ this month ]
Date: Sat, 02 Mar 2024 01:05:53 UTC
Doug Hardie: > > i've done plenty of desktop development work on an M1 Mac (mostly C++ > > and C# projects) and while the performance of the base model isn't > > amazing, it's perfectly usable. on the RPi, on the other hand, i find > > the CPU is too slow to even compile ports usefully[0]. > > That is not surprising. cc uses a lot of disk I/O, and that little > microSD is extremely slow. Replace it with an external SSD and move > everything to it. this is not just about disk I/O; the M1 CPU is simply much, much faster than the RPi4's CPU. to provide some concrete numbers, i did a quick benchmark using primegen[0]. on my desktop, an AMD Ryzen 7 5800X3D (amd64) running 15.0: % time ./primespeed 10000000000 >/dev/null ./primespeed 10000000000 > /dev/null 3.66s user 0.00s system 99% cpu 3.679 total on an RPi 4 running 14.0-RELEASE-p3: % time ./primespeed 10000000000 >/dev/null ./primespeed 10000000000 > /dev/null 21.65s user 0.00s system 99% cpu 21.647 total on a base (8-core) M1 Mac Mini running Darwin 23.2.0: % time ./primespeed 10000000000 >/dev/null ./primespeed 10000000000 > /dev/null 4.27s user 0.00s system 96% cpu 4.422 total this is a single-core benchmark, so we can conclude that an M1 performance core is nearly as fast as a Zen 3 core, while that same M1 code is 5 times faster than an RPi 4 core. and bear in mind that in addition to its 4 performance cores, the base M1 has another 4 efficiency cores. again, my point is not to put down the RPi, which is a great piece of hardware for what it is -- only to point out that, performance-wise, the RPi and the Apple CPUs are really not comparable. regards, lexi. [0] https://cr.yp.to/primegen.html