Interesting analysis of linux kernel threading by IBM

[Larry McVoy]

I'll bet you that the proposed scheduler changes wouldn't have made
the load drop by even 2%, let alone the 600%+ that you got from your
coprocessors.  It sounds like you were completely CPU bound.  

I've been doing kernel performance work for more than ten years, so I'm
always ready to be surprised - in fact, I like it when I'm surprised 
because it means I just learned something.  But I'd be pretty shocked
if the scheduler had any impact whatsoever on your situation.  The only
sort of way that it could is if you were context switching on every
byte.  It's pretty hard to believe that that was the case, because
then you'd have the same problem with your coprocessors.

: Shortly before I went and purchased $10,000 encryption co-processors for
: my SSL web servers it was not unusual to see a 5 min loadave of 30-50 (and
: one time I saw it go up to 144!!). This was on AIX quad 233 2G ram RS/6000
: servers, after I got the encryption co-processors the loadave dropped to
: 5-10
: 
: David Lang 
: 
: On Thu, 20 Jan 2000, Larry McVoy wrote:
: > : need it NOW.  However, the average number of programs people are
: > : running on their machine are increasing, not decreasing.  
: > 
: > This is a completely unsubstatiated statement.  OK, everyone, do this:
: > 
: > $ vmstat
: > load free cach swap pgin  pgou  dk0 dk1 dk2 dk3 ipkt opkt  int  ctx  usr sys idl
: > 0.00  4.3 96.9  0.4    0     0    0   0   0   0    0    0  104   53    0   3  97
: > 
: > Those people with the load field higher than 10, please tell us what you are
: > doing.
: > 
: > -
: > To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
: > the body of a message to majordomo en vger.rutgers.edu
: > Please read the FAQ at http://www.tux.org/lkml/
: > 

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo en vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/