Re: HP35s Sorting Routine Message #28 Posted by Drago Pejic on 30 Aug 2007, 7:37 a.m., in response to message #1 by Nenad (Croatia)
Without any intension to make any commercial advertisement or expose a big ego, I am sure that Data Integrity Institute Inc. has developed and owns the world fastest and most efficient sort algorithm. It is non comparison sort, something very remotely related to radix sort, but it is not radix sort or any academically described.
Remember all quick sorts are so called comparison sorts where you have two kinds of comparisons:
- active comparison, where each item is compared (of course respecting the item type, like integer, float, string, etc.) to others, sometime several times
- passive comparison, where the item position (which is always an integer) in the array and eventually in the file is compared and check that is inside boundary
It is apparently that any comparison will tremendously slow down sorting process. Data Integrity Institute Inc. sorting algorithm does not perform either active (item) comparison or passive (position of the item) one. That allows Data Integrity Institute Inc. Copula Application Specific Enterprise ETL Appliances to process any, does no matter how big and complex data vault, within 3 hours.
Data Integrity Institute Inc. sorting algorithm requires least 512 MB of memory per CPU core in the 32 bit environment and least 128 GB of memory per CPU core in the 64 bit environment. If there is no such memory in the 64 bit environment, efficiency of 64 bit will be several times better than 32 bit, but not such significantly as it could be if there is enough memory available, when it really blasts.
I would like to discuss this, but I am not sure if it is appropriate here. I do not want distract anybody. Also, I can not disclose any secret detail about Data Integrity Institute Inc. sorting algorithm, because it is our most valuable asset and competitive advantage. I can tell that I had two major revelations, something like religious bliss, in ten years while I was obsessed with sorting problem. One was when I went behind quick sort and comparison, and another when I discover how to handle resources in an perfect arrange, which is major obstacle why is radix sort only an academic category.
To Nenad from Croatia:
Puno pozdrava Nenade.
Drago Pejic, M. Sc., F.L.M.I.
http://www.dataintegrityinstitute.com
Edited: 30 Aug 2007, 8:03 a.m.
|