Balancing is never called O(1) in the article. I guess it was somewhat unclear but the scheduler itself is O(1) while load balancing is not constant. I guess it really depends on where you drawn the line between the scheduler and the rest of the kernel.
http://67.163.90.104/~jim/scheduler.html
Thats the completely unedited first draft its all I could find. That has zero editing done, and is even missing a quote up at the top.
It goes further in depth, but as you can see...my unedited writing isn't so great;)
The original article was MUCH MUCH more in depth it was toned down because of the "glaze over" effect people were getting. I didn't have much say in the editorial process in the matter.
I think this way atleast it appeals to a broader audience, not just the CS audience who would know what a binary min heap is if I were to use it.
Slashdot Mirror
Balancing is never called O(1) in the article. I guess it was somewhat unclear but the scheduler itself is O(1) while load balancing is not constant. I guess it really depends on where you drawn the line between the scheduler and the rest of the kernel.
http://67.163.90.104/~jim/scheduler.html Thats the completely unedited first draft its all I could find. That has zero editing done, and is even missing a quote up at the top. It goes further in depth, but as you can see...my unedited writing isn't so great ;)
Sure let me see if I have it around somewhere, the writing is of much lower quality (I'm a programmer after all) but it goes further in.
I'm the author of that article..
The original article was MUCH MUCH more in depth it was toned down because of the "glaze over" effect people were getting. I didn't have much say in the editorial process in the matter.
I think this way atleast it appeals to a broader audience, not just the CS audience who would know what a binary min heap is if I were to use it.
To each his own I suppose.