Kernel Sizes

[Dick Dunn]

gary at sci34hub.sci.com (Gary Heston) writes:
> In article <545 at jahangir.UUCP> marc at jahangir.UUCP (Marc Rossner) writes:
> >I also thought that I had a rather massive kernel on my ISC 2.2.  My
> >kernel is 750K.  Are you sure you have your numbers right?

> My ISC 1.0.6 MultiBus kernel is 573K. Sounds like it's growing
> by leaps and bounds....

Be careful to compare kernels providing the same facilities!  Add a few
device drivers, particularly if they're not well-written, and kernel size
goes through the roof.

> Has anyone surveyed kernel sizes, or graphed the size since UNIX 
> was released? Might me some interesting data in there....

It's all over the map.  Neglecting older kernels that are still running,
in order to get a look at where things are "today"...probably the
smallest "modern" kernel is 10th Edition, in Bell Labs.  (Imagine that! 
The folks who did the original UNIX work still have their act together!)
My guess for the next step up would be BSD kernels with little additional
adornment, in the couple-hundred-Kb range.  The V.3 kernels in 386-land run
1/2-3/4 Mb code, typically around a meg total.  There are some really out-
rageous multi-megabyte kernels for a couple of larger workstation-type
machines.  The trend is definitely to larger kernels, but the range of
sizes has spread a lot:  Within Bell Labs, there's been about a factor of
two increase in the last decade; the other extreme is probably...hmmm...
pushing a factor of a hundred in a decade.  (yuk...what a disgusting
thought...i'm going home)
-- 
Dick Dunn     rcd at ico.isc.com -or- ico!rcd       Boulder, CO   (303)449-2870
   ...If you plant ice, you're gonna harvest wind.