Port-vax archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: netbsd on vax 11/730; booting in sim



Johnny Billquist wrote:
Masao Uebayashi wrote:
True. But noone seems to. And the rest of NetBSD is running along a path
that is more or less causing NetBSD/vax to be unusable anyway.

Just curious - what part is causing so and how?

First of all, gcc is a drag. It consumes so much memory and cpu resources to get anything done that it's not really useful on a VAX anymore. Second, all the dynamic libraries, shared stuff, and all the playing that have been introduced just because it now can be done have also contributed to slowing things down, and that is very noticeable on a slow machine. I'm thinking pam, multibyte support, localization, and so on... Third, we've moved from some simple (but perhaps not always best) programs to way larger and slower alternatives. The most obvious one is replacing sendmail with postfix. I don't think people in general have any idea how long it takes for postfix to even start on a VAX.

Actually, anything that uses c++ is almost guaranteed to be something you don't want. It takes ages to compile, and it takes ages to execute.

I suspect the overhead in file systems, system calls and kernel in general have also grown. There are so many things that people just think are neat and nice, that just aren't that acceptable if you have a slow machine.

Much of all of this is also (atleast) partially true for embedded systems.

Just noticed another "step back", that I hadn't seen until now.

NetBSD moved away quite a long time ago from needing to reboot if fsck needed to fix the root partition. That's something that was done back in the dark ages with 4.3...
But 5.0 have reintroduced this "feature".
Now, I assume it was for some good reason, but I didn't know about it, and as far as I could tell, it worked very nice without having to do this in the last ten years or more.

        Johnny



Home | Main Index | Thread Index | Old Index