Interesting analysis of linux kernel threading by IBM

[Horst von Brand]

"Davide Libenzi" <davidel en maticad.it> said:

[...]

> have You seen the current schedule() implementation ?  It's filled of
> goto to keep a linear CPU fast-path with blocks with lower probability of
> execution kept in tail of the function block.
> To gain how in percent ?
> My patch, IMVHO is more clear then goto jump-back solution, under normal
> cases gives the same results, while under I load gives much better
> performance.

You yourself said it is worse for up to 12 (or was it 8?) running tasks. As
has been shown here convincingly, that is exactly the range in which even
very loaded machines run. So it is a net loss almost everywhere.

[...]

> Remeber that not always strategic decisions are taken by cultured ( in a
> computer science sense ) persons,

Too true.

> otherwise Apple would defeat MSDOS and OS2 would defeat Windows 3.0.

There are other factors. Apple is better by most measures, but absurdly
more expensive. And it shares with OS/2 the absolute absence of
applications. And cost and application availability are crucial. MS is in
the business of getting out a pre-alpha prototype that gets some job done
(sort of), so they catch customers. Then they screw them over with
"upgrades" that don't upgrade anything, in the hope that the next version
will finally work. Or just make a few incompatible changes, so "upgrading"
spreads like a disease. I still remember the pre-MS era, when all kinds of
packages where compared on charts with "number of features", where having
100 fonts in your wodprocessor was somehow better than having 15 (nobody
used more than 3 or 4 anyway). Yep, the market is not about technical
superiority.

But the strength of Linux in exactly that marketplace is technical
superiority, as it just stands outside and doesn't have to go chasing for
high marks in ridiculous benchmarks.
-- 
Horst von Brand                             vonbrand en sleipnir.valparaiso.cl
Casilla 9G, Viña del Mar, Chile                               +56 32 672616

-
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/