|
SoCoder -> Snippet Home -> Variables |
|
|
|
|
Blitz3Dman
|
Root() functionMuch more powerful than Blitz Sqr() function |
|
thanks to this website I was finally able to write this function I've been wondering how to write for a long long time. It's just like Sqr() except it can find any root. (3rd root 4th root 5th root skip a few roots 99th root 100th root etc.) |
|
|
|
|
Comments |
|
| Saturday, 14 April 2007, 15:21 | |
JL235
|
Or you could just use the power sign, which is far faster when on it's own (I get around a factor of 4). |
| Saturday, 14 April 2007, 16:07 | |
Blitz3Dman
|
Oh.. oh well, at least I had the pride of making one anyway 
|
| Friday, 20 April 2007, 08:33 | |
|
Paul |
Actually I'm finding the first code slightly faster. about 5% Don't know if it's the division slowing it down. Nice work both of you
|
| Friday, 20 April 2007, 11:15 | |
|
JL235
|
I get quite a varying range of results. Wrapped in a function, mine might be slightly slower, maybe around 5%. But I'd say that 5% is too close or small to definately say it's slower. It may vary slightly from machine to machine. What is odd, is that outside a function, with both lines on their own, I find mine is almost twice as fast. What should also be noted is that I'd never bother wrapping my snippit in a function, but I would with Blitz3DMan's formula. However as wrapping it in a function makes the most significant difference in performance, I'd say the performance difference must be so neglegable that it wouldn't matter in the vast majority of cases. |
| Saturday, 28 April 2007, 14:07 | |
|
Blitz3Dman
|
Weee! I'm reading my C++ book again and I got to where it tells you how to code functions so I remade my root function in C++ for anybody who cares unfortunately I don't know what the ^ would become in C++. (requires the include math.h) |
| Saturday, 28 April 2007, 23:30 | |
|
hyruleknight |
well the ^ would be the same. i have only seen the mod command change in languages and that was just to %. same say of doing it just change mod with %
|
| Sunday, 29 April 2007, 05:49 | |
|
Blitz3Dman
|
Using both ^ and % the compiler tells me  invalid operands of types `float' and `double' to binary `operator%' |
| Saturday, 23 January 2010, 10:54 | |
|
Zakk |
I've seen ** used instead of ^, it was very odd to me.
|
| Saturday, 23 January 2010, 11:26 | |
Phoenix
|
C++ does not have a power operator. ^ is the xor operator; use the pow function from the standard library for powers. Any inherent speed differences between the two methods are negligible. Performance bottlenecks arise from other causes. |
| Sunday, 24 January 2010, 19:43 | |
|
mindstorm8191 |
JL, your code may or may not be faster, but from what I understand, computers convert those expressions to Blitz3DMan's example anyways. Your code may look more decent, but if you have something like 2^1.5, that's not something easily evaluated. exp(log(2)*1.5), on the other hand, can be done on pen&paper.
|
| Sunday, 24 January 2010, 20:47 | |
|
JL235
|
mindstorm8191 I disagree. If I was getting a difference in speed then I'd predict that my code and his code were compiling down to different native code.JL, your code may or may not be faster, but from what I understand, computers convert those expressions to Blitz3DMan's example anyways.mindstorm8191 Your code may look more decent, but if you have something like 2^1.5, that's not something easily evaluated. exp(log(2)*1.5), on the other hand, can be done on pen&paper.Maybe, but if I was going to calculate the values by hand I'd use a calculator (or more likely an interactive programming language) instead of pen and paper. In which case mine is shorter to type. But I now believe that the most important aspect is clarity. My code is easier to understand in terms of what is happening. This is more important then you think. |
|
|
| |||
|
| ||||