It makes no difference whatsoever. Take the lat_ctx benchmark and make the
processes realtime and see what happens. OK, that's lame, here are the
results:
Realtime:
"size=0k ovr=4.46
5.42 (5.52 5.47 5.43 5.42 5.42 5.42 5.42 5.41 5.40 5.40 5.39)
5.43 (5.45 5.44 5.44 5.43 5.43 5.43 5.42 5.42 5.42 5.41 5.40)
5.41 (5.43 5.43 5.43 5.42 5.42 5.41 5.41 5.41 5.41 5.41 5.40)
normal:
"size=0k ovr=4.44
4.86 (4.98 4.87 4.87 4.86 4.86 4.86 4.86 4.86 4.85 4.85 4.84)
4.61 (5.01 4.85 4.85 4.85 4.79 4.61 4.56 4.55 4.55 4.54 4.53)
4.65 (4.86 4.85 4.84 4.83 4.66 4.65 4.61 4.55 4.55 4.54 4.54)
The numbers in parens are the results from each sub run, the first
number is the median.
So where's the variance due to realtime?
: >. His defense is completely unreasonable, both from a statistical
: > point of view (consider his standard deviation)
:
: So he needs to run it longer. No problem.
Show me how that helps.
: >. While it would be nice for Richard if someone dissected his benchmark,
: > the people who can have already pruned that branch of the tree as
: > uninteresting, i.e., no payoff.
:
: No payoff UNLESS you run real-time processes.
Been there, tried that, his benchmark is still flawed.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/