Does the magnitude of math.random values affect performance?
The math.random() function has 2 values that the chosen number is between, m and n (as the wiki says).
Example: I have a script that gets a random value for the transparency of some part, that can be between 0 and 1. I can make
math.random(0, 1) -- can return me 2 values (0 or 1)
or
math.random(0, 100) / 100 -- can return me 101 values (like 0.06 or 0.84)
Ignoring the division by 100, does the number of options to randomly choose affect a script performance? Will the computer have to calculate a hundred times more?
(it may sound insignificant, but if I do it in a constant loop, for example, can I feel the difference?)
1 answer
The difference is almost completely negligible. I went ahead and ran a test, and the average difference between math.random(0,1) and math.random(0,100) is 9.5 microseconds (0.0000095 seconds.) You really don't need to worry.
In case you're wondering, this is script I used to figure that out:
local huget = {}
local smallt = {}
local hugeav = 0
local smallav = 0
for i = 1, 100 do
local t = tick()
print(math.random(0, 100))
table.insert(huget,tick()-t)
wait()
end
for i = 1, 100 do
local t = tick()
print(math.random(0,1))
table.insert(smallt,tick()-t)
wait()
end
for i,v in ipairs(huget) do
hugeav = hugeav + v
end
hugeav = hugeav/#huget
for i,v in ipairs(smallt) do
smallav = smallav + v
end
smallav = smallav/#smallt
print("average math.random(0,100) time: " .. hugeav)
print("average math.random(0,1) time: " .. smallav)
0
0