login
Header Space

 
 

CFS performance

November 13, 2007 - 12:19pm
Submitted by Anonymous on November 13, 2007 - 12:19pm.
Linux

Due to its design, the CFS scheduler is not prone to any of the
+ 'attacks' that exist today against the heuristics of the stock
+ scheduler: fiftyp.c, thud.c, chew.c, ring-test.c, massive_intr.c all
+ work fine and do not impact interactivity and produce the expected
+ behavior

This is what Ingo Molnar claims in his CFS patch.
what are these files and attacks? How can I test CFS against these attacks? Where are these files available?

Here is one of the above

November 13, 2007 - 12:44pm

Here is one of the above examples. Search the kernel archives on KT, or google for them. They aren't that difficult to find.

These "attacks" are merely ways that people have found to chew up more CPU time than you would normally expect the running program to be given. Someone that can break into your system to compile these, slowing your system, is probably more likely to do something much more interesting/devastating than that. The so-called attacks are really nothing to get excited about.

That being said the CFS seems to be a vast improvement in the scheduler from what previously was used, and is worth getting.

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.
speck-geostationary