I have two tasks:
val t = System.currentTimeMillis()
def first: fs2.Task[Int] = fs2.Task.fromFuture(Future.successful {
println(s"${ System.currentTimeMillis() - t} - First")
Thread.sleep(100)
1
})
def second: fs2.Task[Int] = fs2.Task.fromFuture(Future.successful {
println(s"${System.currentTimeMillis() - t} - Second")
Thread.sleep(200)
2
})
I want to start them in parallel, and run a function on results in the end (pretty much as I'd do vs Applicative map2).
This works perfectly fine:
val comb = for {
ta ← fs2.Task.start(first)
tb ← fs2.Task.start(second)
a ← ta
b ← tb
} yield {
println("Complete")
a + b
}
comb.unsafeRun()
output: 164 - First
164 - Second
It is a bit tedious to write extra line in for comprehension every time.
To my suprise neither
import fs2.interop.scalaz._
import scalaz.syntax.applicative._
val comb2 = (first |#| second) {_ + _}
output: 197 - First 318 - Second
nor
val comb1 = for {
a ← first.async
b ← second.async
} yield {
a + b
}
output: 202 - First 323 - Second
worked.
While in the first case tasks started simultaneuosly, in the latter 2, they began one after another.
What am I missing?
Reading the implementation of async I understood, that intent is to make first snippet syntactically attractive, but I cant believe that there is a bug, so I possibly just misuse async?
Related
If I have (for example) a table with 300 entries, how would I get the latest x entries only?
I was thinking of doing the next, but I'm wondering if there is a better/more optimized way to do this exact thing.
local TestTable = {}
-- Populate table
for i = 1, 300, 1 do
print('Adding: ' .. i)
table.insert(TestTable , i)
end
-- Get latest x of table
function GetLatestFromTable(OriginalTable, Amount)
local TableLength = #OriginalTable
local Retval = {}
for i = 1, Amount, 1 do
if TableLength - i <= 0 then break end -- Dont allow to go under 0
table.insert(Retval, OriginalTable[TableLength - i])
print("Adding to Retval: " .. OriginalTable[TableLength - i] .. ' (Index: ' .. TableLength - i .. ')')
end
return Retval
end
print(#TestTable)
local LatestTable = GetLatestFromTable(TestTable, 10)
print(#LatestTable)
For keys in sequence (and values are string/number) a call to table.concat() allows range parameter.
local tab = {"One", "Two", "Three", "Four", "Five"}
print(table.concat(tab, '\n', #tab - 1, #tab)) -- Last two entries
See: table.concat()
As mentioned by #Luke100000, one way could be to use Lua custom iterators. In Lua, an iterator is a special function which, when called, will return the next value. It is made possible by the fact that functions are first-class citizen in Lua and they can refer to previous scope with a mecanism named closure.
To answer the question, one could start implement a general iterator over a given range.
function IterateRange (Table, Min, Max)
local ClosureIndex = Min - 1
local ClosureMax = math.min(Max, #Table)
local function Closure ()
if (ClosureIndex < ClosureMax) then
ClosureIndex = ClosureIndex + 1
return Table[ClosureIndex]
end
end
return Closure
end
IterateRange is a function returning an anonymous function. The anonymous function does not take any parameter. It simply update the ClosureIndex index defined in the local scope of IterateRange and return the table value.
The first thing that the anonymous function do is to increment ClosureIndex. For that reason, ClosureIndex must be initialized to Min - 1.
This function works as one might expect:
TestTable = {}
for i = 1, 300, 1 do
print('Adding: ' .. i)
table.insert(TestTable , i)
end
for Value in IterateRange(TestTable, 290, 300) do
print(Value)
end
290
291
292
293
294
295
296
297
298
299
300
Now, it's trivial to reuse this general iterator to iterate over the last N entries of a given table:
function IterateLastEntries (Table, Count)
local TableSize = #Table
local StartIndex = (TableSize - Count)
return IterateRange(Table, StartIndex, TableSize)
end
It also work as one might expect:
TestTable = {}
for i = 1, 300, 1 do
print('Adding: ' .. i)
table.insert(TestTable , i)
end
for Value in IterateLastEntries(TestTable, 10) do
print(Value)
end
290
291
292
293
294
295
296
297
298
299
300
And finally, to summarize all this in a fully copy & pasteable solution:
TestTable = {}
for i = 1, 300, 1 do
print('Adding: ' .. i)
table.insert(TestTable , i)
end
function IterateRange (Table, Min, Max)
local ClosureIndex = Min - 1
local ClosureMax = math.min(Max, #Table)
local function Closure ()
if (ClosureIndex < ClosureMax) then
ClosureIndex = ClosureIndex + 1
return Table[ClosureIndex]
end
end
return Closure
end
function IterateLastEntries (Table, Count)
local TableSize = #Table
local StartIndex = (TableSize - Count)
return IterateRange(Table, StartIndex, TableSize)
end
for Value in IterateLastEntries(TestTable, 10) do
print(Value)
end
This should return:
290
291
292
293
294
295
296
297
298
299
300
I will let the OP update the code in order to achieve the same results for 30 entries.
var n = 27;
void main() {
while (n>=5)
if ( (n) % 3==0 && (n) % 2==0 ){
print(n-1);
break;
}
else if ( (n) % 3==0 && (n) % 2!=0 ){
print(n-2);
break;
}
else if ((n) % 3!=0 && (n) % 2==0){
print((n/6).floor()*6 +(1));
break;
}
else { n=n+1;}
}
I was looking for a way to have an input reduced to the nearest 1 or 5mod6 (for n>=5) and came up with the code above in Dart (project is in flutter). Does anyone have any better way of doing the same thing?
The result (rounded down input) will then be passed to another function.
For the current n value of 27 the console will print 25... try other values 24 maps to 23, 23 to itself, 22,21,20 to 19, 19 to itself, 18 to 17, 17 to itself, 16,15,14 to 13.....and I hope you get the idea.
I would get rid of the loop by finding the next lower multiple of 6 and then adding 1 or 5. The code ends up being shorter, and I think it expresses your intent much more clearly than code involving % 3 and % 2.
int lowerMultiple(int n, int multipleOf) =>
((n - 1) / multipleOf).floor() * multipleOf;
void main() {
var n = 27;
var m = lowerMultiple(n, 6);
print((n - m) >= 5 ? (m + 5) : (m + 1));
}
The above code should work for integers less than 5 as well, including non-positive ones.
Suppose I write this function
function test_function(T)
c = 1
d = 31
q = 321
b = 32121
a = 10
for i in 1:T
c = d + q + b + a
end
end
There will be no memory allocation. However, in my own code, I wrote a similar loop, but I encounter a huge amount of memory allocation. I can't share the entirety of my code, but when I used --track-allocation=user, I see the following results
80000 q = 3
- p = 0.1
- p_2 = 3
- q_2 = .2
-
240000 r = p - p_2 + q_2 - q;
The code above is in a for loop. This is just strange to me - why would Julia ever allocate memory to single digits?
I can't understand why this one works:
var arr = [4,5,6,7]
arr.map() {
x in
return x + 2
}
while this one not
arr.map() {
x in
var y = x + 2
return y
}
with error
Playground execution failed: MyPlayground.playground:13:5: error:
cannot invoke 'map' with an argument list of type '((_) -> _)'
arr.map() {
The problem here is there error message. In general, when you see something like cannot invoke .. with ... it means that the compiler's type inference has just not worked.
In this case, you've run up against one of the limitations of inference within closures. Swift can infer the type of single-statement closures only, not multiple-statement ones. In your first example:
arr.map() {
x in
return x + 2
}
There's actually only one statement: return x + 2. However, in the second:
arr.map() {
x in
var y = x + 2
return y
}
There's an assignment statement (var y = x + 2), and then the return. So the error is a little misleading: it doesn't mean you "can't invoke map() with this type of argument", what it means to say is "I can't figure out what type x or y is".
By the way, in single-statement closures, there are two other things that can be inferred. The return statement:
arr.map() {
x in
x + 2
}
And the variable name itself:
arr.map() { $0 + 2 }
It all produces the same compiled code, though. So it's really a matter of taste which one you choose. (For instance, while I think the inferred return looks clean and easier to read, I don't like the $0, so I generally always put x in or something, even for very short closures. It's up to you, though, obviously.)
One final thing: since this is all really just syntax stuff, it's worth noting that the () isn't needed either:
arr.map { x in x + 2 }
As #MartinR pointed out, the compiler can infer some types from outer context as well:
let b: [Int] = arr.map { x in
var y = x + 2
return y
}
Which is worth bearing in mind. (it seems that the "one-statement" rule only applies when there's no other type info available)
Swift can't infer type every time. Even though it should see that y = x + 2 means y is an Int too. My guess is that Swift parses the closure in a certain order that makes it not aware of the return type ahead of time in your case.
This works:
arr.map() {
x -> Int in
var y = x + 2
return y
}
I've been struggling with the following code. It's an F# implementation of the Forward-Euler algorithm used for modelling stars moving in a gravitational field.
let force (b1:Body) (b2:Body) =
let r = (b2.Position - b1.Position)
let rm = (float32)r.MagnitudeSquared + softeningLengthSquared
if (b1 = b2) then
VectorFloat.Zero
else
r * (b1.Mass * b2.Mass) / (Math.Sqrt((float)rm) * (float)rm)
member this.Integrate(dT, (bodies:Body[])) =
for i = 0 to bodies.Length - 1 do
for j = (i + 1) to bodies.Length - 1 do
let f = force bodies.[i] bodies.[j]
bodies.[i].Acceleration <- bodies.[i].Acceleration + (f / bodies.[i].Mass)
bodies.[j].Acceleration <- bodies.[j].Acceleration - (f / bodies.[j].Mass)
bodies.[i].Position <- bodies.[i].Position + bodies.[i].Velocity * dT
bodies.[i].Velocity <- bodies.[i].Velocity + bodies.[i].Acceleration * dT
While this works it isn't exactly "functional". It also suffers from horrible performance, it's 2.5 times slower than the equivalent c# code. bodies is an array of structs of type Body.
The thing I'm struggling with is that force() is an expensive function so usually you calculate it once for each pair and rely on the fact that Fij = -Fji. But this really messes up any loop unfolding etc.
Suggestions gratefully received! No this isn't homework...
Thanks,
Ade
UPDATED: To clarify Body and VectorFloat are defined as C# structs. This is because the program interops between F#/C# and C++/CLI. Eventually I'm going to get the code up on BitBucket but it's a work in progress I have some issues to sort out before I can put it up.
[StructLayout(LayoutKind.Sequential)]
public struct Body
{
public VectorFloat Position;
public float Size;
public uint Color;
public VectorFloat Velocity;
public VectorFloat Acceleration;
'''
}
[StructLayout(LayoutKind.Sequential)]
public partial struct VectorFloat
{
public System.Single X { get; set; }
public System.Single Y { get; set; }
public System.Single Z { get; set; }
}
The vector defines the sort of operators you'd expect for a standard Vector class. You could probably use the Vector3D class from the .NET framework for this case (I'm actually investigating cutting over to it).
UPDATE 2: Improved code based on the first two replies below:
for i = 0 to bodies.Length - 1 do
for j = (i + 1) to bodies.Length - 1 do
let r = ( bodies.[j].Position - bodies.[i].Position)
let rm = (float32)r.MagnitudeSquared + softeningLengthSquared
let f = r / (Math.Sqrt((float)rm) * (float)rm)
bodies.[i].Acceleration <- bodies.[i].Acceleration + (f * bodies.[j].Mass)
bodies.[j].Acceleration <- bodies.[j].Acceleration - (f * bodies.[i].Mass)
bodies.[i].Position <- bodies.[i].Position + bodies.[i].Velocity * dT
bodies.[i].Velocity <- bodies.[i].Velocity + bodies.[i].Acceleration * dT
The branch in the force function to cover the b1 == b2 case is the worst offender. You do't need this if softeningLength is always non-zero, even if it's very small (Epsilon). This optimization was in the C# code but not the F# version (doh!).
Math.Pow(x, -1.5) seems to be a lot slower than 1/ (Math.Sqrt(x) * x). Essentially this algorithm is slightly odd in that it's perfromance is dictated by the cost of this one step.
Moving the force calculation inline and getting rid of some divides also gives some improvement, but the performance was really being killed by the branching and is dominated by the cost of Sqrt.
WRT using classes over structs: There are cases (CUDA and native C++ implementations of this code and a DX9 renderer) where I need to get the array of bodies into unmanaged code or onto a GPU. In these scenarios being able to memcpy a contiguous block of memory seems like the way to go. Not something I'd get from an array of class Body.
I'm not sure if it's wise to rewrite this code in a functional style. I've seen some attempts to write pair interaction calculations in a functional manner and each one of them was harder to follow than two nested loops.
Before looking at structs vs. classes (I'm sure someone else has something smart to say about this), maybe you can try optimizing the calculation itself?
You're calculating two acceleration deltas, let's call them dAi and dAj:
dAi = r*m1*m2/(rm*sqrt(rm)) / m1
dAj = r*m1*m2/(rm*sqrt(rm)) / m2
[note: m1 = bodies.[i].mass, m2=bodies.[j].mass]]
The division by mass cancels out like this:
dAi = rm2 / (rmsqrt(rm))
dAj = rm1 / (rmsqrt(rm))
Now you only have to calculate r/(rmsqrt(rm)) for each pair (i,j).
This can be optimized further, because 1/(rmsqrt(rm)) = 1/(rm^1.5) = rm^-1.5, so if you let r' = r * (rm ** -1.5), then Edit: no it can't, that's premature optimization talking right there (see comment). Calculating r' = 1.0 / (r * sqrt r) is fastest.
dAi = m2 * r'
dAj = m1 * r'
Your code would then become something like
member this.Integrate(dT, (bodies:Body[])) =
for i = 0 to bodies.Length - 1 do
for j = (i + 1) to bodies.Length - 1 do
let r = (b2.Position - b1.Position)
let rm = (float32)r.MagnitudeSquared + softeningLengthSquared
let r' = r * (rm ** -1.5)
bodies.[i].Acceleration <- bodies.[i].Acceleration + r' * bodies.[j].Mass
bodies.[j].Acceleration <- bodies.[j].Acceleration - r' * bodies.[i].Mass
bodies.[i].Position <- bodies.[i].Position + bodies.[i].Velocity * dT
bodies.[i].Velocity <- bodies.[i].Velocity + bodies.[i].Acceleration * dT
Look, ma, no more divisions!
Warning: untested code. Try at your own risk.
I'd like to play arround with your code, but it's difficult since the definition of Body and FloatVector is missing and they also seem to be missing from the orginal blog post you point to.
I'd hazard a guess that you could improve your performance and rewrite in a more functional style using F#'s lazy computations:
http://msdn.microsoft.com/en-us/library/dd233247(VS.100).aspx
The idea is fairly simple you wrap any expensive computation that could be repeatedly calculated in a lazy ( ... ) expression then you can force the computation as many times as you like and it will only ever be calculated once.