Coding Challenge - Vampire Numbers Challenge - Results

Challenge Description:

The goal of this Challenge was to build an application that returns all Vampire Numbers within a given range. A vampire number is an integer

with an even number of digits
with a pair of factors (called the fangs) having two properties:
1. they have an equal number of digits
2. together, they possess exactly the same digits as the vampire number
3. not gotten by simply adding zeros to the end of a smaller vampire number

The submissions for this contest were judged on three criteria:

The first was accuracy. The VI must successfully find all of the Vampire Numbers in a given range. This could be integers with a small number of digits, or integers with more digits than many data types can hold. The next criterion was speed. For those submissions that did find all Vampire Numbers in a given range, how fast could they find these? Finally, and very importantly, submissions were judged on code readability. Even if a VI performs a specific task, it must be readable by other programmers and follow LabVIEW Style Guidelines.

Coding Challenge Winner

Christian Altenbach, from UCLA, is the Vampire Number Coding Challenge winner! His submission efficiently returned Vampire Numbers in wide ranges of numbers and also displayed good LabVIEW programming style.

Winning Entry Description:

1. Accuracy: To determine Vampire Numbers which contain more digits than a simple integer will hold, Christian devised the following plan. In his application, the raw "beginning" and "end" boundaries are used to generate individual ranges. All parts with an odd number of digits or less than 4 digits are eliminated. Each range with an equal number of digits is treated separately. The digit count determines which version of basically the same search code is executed:

4-8 digits: I32 based
10-14 digits: DBL based
16 digits: EXT based
18+ digits: Unf (Unsigned integer of infinite precision) based, represented as array of digits.)

He also built set of Unf math tools (multiply, add, increment, etc.) that make certain assumptions about the data. And in these VIs, Unf multiplication is Fourier based (base 10).

By developing this structure, Christian's VI found the correct amount of Vampire Numbers in each of the ranges tested. They were as follows:

Range of Numbers Tested	# Vampire Numbers Found
1 - 10,000	7
1 - 1,000,000	155
1 - 100,000,000	3383
0 - 999,999,999	5184
1000000000-1001000000	9
900000000000-900020000000	3
99900000000000-99999999999999	76
10000000000000-10000100000000	17

2. Speed: All submissions were run on a Dell Dimension computer with a 2.4 GHz Pentium 4 processor and 512MB RAM. Christian's VI was the only submission which found all Vampire Numbers for each range tested in reasonable time. His submission completed the ranges in the specified times:

Range of Numbers Tested	Time to Find all Vampire Numbers
1 - 10,000	1 millisecond
1 - 1,000,000	55 milliseconds
1 - 100,000,000	5.2 seconds
0 - 999,999,999	5.2 seconds
1000000000-1001000000	458 milliseconds
900000000000-900020000000	480 milliseconds
99900000000000-99999999999999	9.5 seconds
10000000000000-10000100000000	47.7 seconds

Honorable Mention:
Heiko Fettig, from Bookham Technology, also submitted an excellent solution. Although it did not find Vampire Numbers in all ranges within reasonable time, it performed extremely well, sometimes faster, in others. Heiko's VI successfully found the correct Vampire Numbers in the following ranges within the listed time:

Range of Numbers Tested	Time to Find all Vampire Numbers
Range of Numbers Tested	Time to Find all Vampire Numbers
1 - 10,000	9 milliseconds
1 - 1,000,000	42.6 milliseconds
0 - 999,999,999	4.1 seconds

3. Coding Style: Even when a VI correctly performs a specific task, it is also important that the VI be readable by other developers. Christian's VI was coded very well

Front Panel
- Well organized
- Contained to one font
- Object evenly aligned and distributed

Block Diagram

Clean wiring techniques (very important!)
Efficient memory management (minimization of data copies)
Very good use of comments
Though top-level VI Block Diagram was larger than screen size, it was only necessary to scroll in one direction

enlarge

Icon and Connector Pane

Very unique, descriptive icons

No colloquialisms on icons

Honorable Mention:
Heiko Fettig, from Bookham Technology, also rated very well in this category. His Block Diagrams were very clean and, again, he created come excellent icons for his application. The entire Block Diagram for his top-level VI is below:

Notice the clear user of comments and compact nature of his Block Diagram. It is easy to follow and clean of messy wires.

Again, thank you to everyone who competed in this competition. And congratulations to Christian Altenbach, the overall winner for the Vampire Numbers Coding Challenge. Be sure to check out the current Coding Challenge for even more LabVIEW fun!