Return Styles: Pseud0ch, Terminal, Valhalla, NES, Geocities, Blue Moon.

Pages: 1-

You're doing it wrong.

Name: Anonymous 2010-06-14 22:24

http://queue.acm.org/detail.cfm?id=1814327
Did you just decide that my order of magnitude claim was bogus, because it is based on only an extreme corner case? If so, you are doing it wrong, because this is pretty much the real-world behavior seen.
So true.

Name: Anonymous 2010-06-14 22:29

>>1
Theoretical computer scientists have been doing practical server optimization wrong for decades? Oh golly.
Obviously this idiot is oblivious to the fact there is no beauty or greater understanding to come out of trivially optimizing a heap implementation for your particular operating system. Advancements to the field don't come from bit twiddling and monitoring cache-misses, this is pathetic.

Name: Anonymous 2010-06-14 22:36

What are you, some kind of robot?

Name: Anonymous 2010-06-14 22:48

A factor of 10 is smalltime and leaves you in the same O(). This is just a microoptimization, and a pretty old one at that.

Name: Anonymous 2010-06-14 22:50

>>4
What's cheaper a server or ten servers? I'd say such an optimization is good enough for the real world.

Name: Anonymous 2010-06-14 22:54

>>5
So it may be, but to pretend like anyone in academia would or even should give a shit is pure ignorance.

Name: Anonymous 2010-06-14 23:04

>>6
Then academia is irrelevant.

Name: Anonymous 2010-06-14 23:08

>>4
I don't think you read the article or understand big-O. His gives you a worse O() but better performance. That's the point: O(log(n)) doesn't mean much when you have constant extra-algorithmic interruptions1, but if using a different algorithm eliminates most of them the story changes drastically.

Oh and if you're server is swapping, well, your doing it wrong too.
IHBT! Also you're fired.

_____________________________
1: Suppose you even had O(1) -- almost all of your time would be spent going to disk which is not captured by your big-O figure.

Name: Anonymous 2010-06-14 23:16

>>7
TOLD

Name: Anonymous 2010-06-15 0:31

>>7,9
SPAWHBT.

Name: Anonymous 2010-06-15 0:53

>>7
In case you are serious- the job of computer scientists is not to figure out how to best write algorithms to conform to specific memory models (Although this is an area of research for some), but to find algorithms and data structures that are intrinsically "better" by some common common measure than others.

This scenario is not very dissimilar to me designing a special computer that runs bubble sort faster than merge sort; then proceeding to mock academics for discarding bubble sort, when it  is clearly superior.

Name: Anonymous 2010-06-15 4:46

Practice often trumps theory, thats one of the way theory advances by leaps and bounds. Consider which of these contributed more to the progress of either of two:
Calculations and Math
Engineering and Science
Programming and Computer Science

Name: Graph Punk !!XGrol9omL3xbWra 2010-06-15 5:23

>>12
All of these things are very closely related.

Name: Anonymous 2010-06-15 10:59

>>12
Calculations & Math: Calculations & Math; Engineering and Science (Both disciplines have been around pretty much forever and have always been closely related)
Engineering & Science: Engineering & Science; Calculations & Math ("")
Programming & Computer Science: Calculations & Math; Programming & Computer science (Programming & Computer Science is basically applied math)

Name: Anonymous 2010-12-17 1:22

Are you GAY?
Are you a NIGGER?
Are you a GAY NIGGER?

If you answered "Yes" to all of the above questions, then GNAA (GAY NIGGER ASSOCIATION OF AMERICA) might be exactly what you've been looking for!

Name: Anonymous 2011-02-04 16:49

Don't change these.
Name: Email:
Entire Thread Thread List
All trademarks and copyrights on this site are owned by their respective parties. All comments and posts are the responsibility of their own posters.
- Tinychan + 2ch Mode -