Kernel Compiling Checklist, rev 3

[Christopher Ness]

On Tue, 2004-04-13 at 11:23, jludwig wrote:
> One more item make -j # > /dev/null allows parallel compiling of
> components where # is the number of forks. I use about 1 per 64 meg
> allowing for overhead, and dumping messages also speeds the process and
> you still get warnings and errors.-- 
> jludwig <wralphie at comcast.net>

Nice tip!  I never thought of redirecting STDOUT to the bit bucket to
gain a speedup while compiling.  Classic Unix flavour too with "no news
is good news".

For the threads, wouldn't you only see a speedup on SMP machines?  Does
kernel compilation have a lot of HDD waiting where another thread can
jump in and use the CPU.

I would think two, maybe three threads would be the max for a single CPU
machine after that the overhead of threading would kill any speedup. 
But you suggest for my laptop -> 512/64 = 8 threads.  How did you come
up with the 64M guideline?

Sorry, lots of questions in there.  :)
Chris
-- 
Software Engineering IV,
McMaster University
PGP Public Key: http://nesser.org/pgp-key/

11:38:23 up 3:18, 1 user, load average: 0.58, 0.58, 0.46 

Please avoid sending me Word or PowerPoint attachments.
See http://www.fsf.org/philosophy/no-word-attachments.html
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
URL: <http://listman.redhat.com/archives/fedora-list/attachments/20040413/9d389a74/attachment-0001.sig>