Sorting over 1 million numbers

Quote:

---

Originally posted by zboss:
Use a Hash Function. This doesn't sort the items at all, but uses pointers. That way when you need to insert/delete a new item into the list, you don't have to move the items. So, the first time around may be a little slower, but subsequent performance will be much faster.

---

I thought we talked about this, using pointers is very expensive in this case. Not a good soln'.

------------------
DHAHL3seasons GP:73 G:121 A:55 Pts:176 GWG:12 +/-:184
UWSWA1season GP:9 G:12 A:8 Pts:20 GWG:3 +/-:-3

uwcdc.com or namgor.com

Quote:

---

Originally posted by sjelkjd:

Quote:

---

Originally posted by EdtheROM:
The underlying disk/file structure hasn't changed much, so the hint to use multiples of the cluster size is a good one.

Just why would you think that virtual memory would obviate this??

I'm interested...

-ed-

---

I suspect that if you look at the number of page accesses for a "dumb" quicksort and

*********
*********
*********

That's why i think that this mergesort is a big loser in this case: you're going to have to go to disk roughly twice as much.

---

*****

given all of this, have you ever considered Singleton's sort??

-ed-

Quote:

---

Originally posted by EdtheROM:
*****

given all of this, have you ever considered Singleton's sort??

-ed-

---

Actually, i've never heard of Singleton's sort.
Could you describe the algorithm?