What is more efficient
if ($n-1 < $x)
or
if ($n <= $x)
Anyone know?
views:
217answers:
6
+6
A:
Probably ($n <= $x) since there's one less operation (no subtraction). However, this is only equivalent for integers and the difference is very likely insignificant.
Artefacto
2010-08-02 18:46:44
If the language you're using doesn't automatically pick the more efficient instructions, then you should log a missed-optimization bug against your language's implementation.
Bernd Jendrissek
2010-08-02 19:49:41
@Bernd Seriously?... I doubt the subtraction could easily be optimized way, especially given that in C, for two ints `n - 1 < x` is different from `n <= x` (the integer can wrap around).
Artefacto
2010-08-03 01:21:26
Yikes! I didn't think of overflow. Then again, two things: the OP is about PHP, and in C the behaviour on overflow is undefined. So a (C) compiler could legitimately transform `n-1<x` into `n<=x`, but not vice versa.
Bernd Jendrissek
2010-08-10 16:20:17
Why is this the most upvoted? this didn't even answer the question correctly, the second one is more efficient.
Shawn Mclean
2010-08-02 19:17:12
The point of my answer is that it doesn't freaking matter AT ALL and you should strongly prefer the clearer syntax. I guess 10 people agreed.
Triptych
2010-08-02 19:20:15
I agree with "just pick the clearer syntax", but disagree with the assertion that that is always the same one of the options. Which one is clearer depends on the application. If you're reserving the first, or last, element of some array, $n-1 < $x might be clearer.
Bernd Jendrissek
2010-08-02 19:48:41
+3
A:
In the first instance you're performing two operations (subtract then compare) whereas in the second you're performing just one (compare). I think it's safe to say the second is more performant. Regardless, the first form is very nonstandard and I don't know why it would ever be used as a substitute for the second.
Jordan
2010-08-02 18:47:33
+4
A:
It shouldn't be significant speed gain but later seems more readable to me. The first one has an extra subtraction part too.
Sarfraz
2010-08-02 18:47:43
+1
A:
Since you're using a high-level scripting language, it's not a matter of processor operations to perform that action. More time will be spent in parsing the code, validating variable types and managing variables memory and many other things.
Even if there's no practical difference, I guess the second will be faster since the language expression is simpler.
The term $n-1 is in a section more complex to understand that <=. The math operator has only a few options (>, <, <=, ==, >=, !=, ...), but the term $n-1 requires a more complex analysis because it has to be split by the minus sign at first, and then analyzed both parts (which could also be a function call, a constant, a variable, another complex expression, ...)
Diego Jancic
2010-08-02 19:06:04
+3
A:
Neither. PHP has no enforced implementation saying weather it's bytecode, interpreted directly from the AST, or compiled to machine code. In fact, there are a billion different implementations and the main one probably changes all the time just like the rest of the language and API.
Even at the assembly level in x86, there is no difference between the two operations. Proof:
x < y
cmp eax, y
jb its_true
x <= y
cmp eax, y
jbe its_true
JBE is a synonym to something like JC IIRC, which just checks a single flag. Both of these conditional branches simply check if a flag is set or not.
The only real way these operations could effect speed is you could overload both less than/greater than operators and equality, and even then the speed difference should be negligible.
Longpoke
2010-08-02 20:02:24