I have the following data:
A = [a0 a1 a2 a3 a4 a5 .... a24]
B = [b0 b1 b2 b3 b4 b5 .... b24]
which I then want to multiply as follows:
C = A * B' = [a0b0 a1b1 a2b2 ... a24b24]
This clearly involves 25 multiplies.
However, in my scenario, only 5 new values are shifted into A per "loop iteration" (and 5 old values are shifted out of A). Is there any fast way to exploit the fact that data is shifting through A rather than being completely new? Ideally I want to minimize the number of multiplication operations (at a cost of perhaps more additions/subtractions/accumulations). I initially thought a systolic array might help, but it doesn't (I think!?)
Update 1: Note B is fixed for long periods, but can be reprogrammed.
Update 2: the shifting of A is like the following: a[24] <= a[19], a[23] <= a[18]... a[1] <= new01, a[0] <= new00. And so on so forth each clock cycle
Many thanks!
Is there any fast way to exploit the fact that data is shifting through A rather than being completely new?
Even though all you're doing is the shifting and adding new elements to A, the products in C will, in general, all be different since one of the operands will generally change after each iteration. If you have additional information about the way the elements of A or B are structured, you could potentially use that structure to reduce the number of multiplications. Barring any such structural considerations, you will have to compute all 25 products each loop.
Ideally I want to minimize the number of multiplication operations (at a cost of perhaps more additions/subtractions/accumulations).
In theory, you can reduce the number of multiplications to 0 by shifting and adding the array elements to simulate multiplication. In practice, this will be slower than a hardware multiplication so you're better off just using any available hardware-based multiplication unless there's some additional, relevant constraint you haven't mentioned.
on the very first 5 data set you could be saving upto 50 multiplications. but after that its a flat road of multiplications. since for every set after the first 5 set you need to multiply with the new set of data.
i'l assume all the arrays are initialized to zero.
i dont think those 50 saved are of any use considering the amount of multiplication on the whole.
But still i will give you a hint on how to save those 50 maybe you could find an extension to it?
1st data set arrived : multiply the first data set in a with each of the data set in b. save all in a, copy only a[0] to a[4] to c. 25 multiplications here.
2nd data set arrived : multiply only a[0] to a[4](having new data) with b[0] to b[4] resp. save in a[0] to a[4],copy to a[0->9] to c. 5 multiplications here
3rd data set arrived : multiply a[0] to a[9] with b[0] to b[9] this time and copy to corresponding a[0->14] to c.10 multiplications here
4th data set : multiply a[0] to a[14] with corresponding b copy corresponding a[0->19] to c. 15 multiplications here.
5th data set : mutiply a[0] to a[19] with corresponding b copy corresponding a[0->24] to c. 20 multiplications here.
total saved mutiplications : 50 multiplications.
6th data set : usual data multiplications. 25 each. this is because for each set in the array a there a new data set avaiable so multiplication is unavoidable.
Can you add another array D to flag the changed/unchanged value in A. Each time you check this array to decide whether to do new multiplications or not.
Related
A person claims that they can improve InsertionSort by the following argument. In the innermost loop of InsertionSort, instead of looping over all entries in the already sorted array in order to insert the j’th observed element, simply perform BinarySearch in order to sandwich the j’th element in its correct position in the list A[1, ... , j−1]. This person claims that their resulting insertion sort is asymptotically as good as mergesort in the worst case scenario. True or False and why? Circle the one correct answer from the below:
a. True: In this version, the while loop will iterate log(n), but in each such iteration elements in the left side of the list have to be shifted to allow room for the key to propagate downwards across the median elements and so this shift will still require log(n) in the worst case scenario. Adding up, Insertion Sort will significantly improve in this case to continue to require n log(n) in the worst case scenario like mergesort.
b. False: In this version, the while loop will iterate log(n), but in each such iteration elements in the left side of the list have to be shifted to allow room for the key to propagate downwards and so this shift will still require n in the worst case scenario. Adding up, Insertion Sort will continue to require n² in the worst case scenario which is orders of magnitude worse than mergesort.
c. False: In this version, the while loop will iterate n, but in each such iteration elements in the left side of the list have to be shifted to allow room for the key to propagate downwards and so this shift will still require log(n) in the worst case scenario. Adding up, Insertion Sort will continue to require n log(n) in the worst case scenario which is orders of magnitude worse than mergesort.
d. True: In this version, the while loop will iterate log(n), but in each such iteration elements in the left side of the list have to be shifted to allow room for the key to propagate downwards and so this shift will still require n in the worst case scenario. Adding up, Insertion Sort will continue to require n log(n) in the worst case scenario which is orders of magnitude worse than mergesort.
b is correct, with some assumptions about compiler optimizations.
Consider a reverse sorted array,
8 7 6 5 4 3 2 1
and that insertion sort is half done so it is
5 6 7 8 4 3 2 1
The next step:
normal insertion sort sequence assuming most recent value read kept in register:
t = a[4] = 4 1 read
compare t and a[3] 1 read
a[4] = a[3] = 8 1 write
compare t and a[2] 1 read
a[3] = a[2] = 7 1 write
compare t and a[1] 1 read
a[2] = a[1] = 6 1 write
compare t and a[0] 1 read
a[1] = a[0] = 5 1 write
a[0] = t = 4 1 write
---------------
5 read 5 write
binary search
t = a[4] 1 read
compare t and a[1] 1 read
compare t and a[0] 1 read
a[4] = a[3] 1 read 1 write
a[3] = a[2] 1 read 1 write
a[2] = a[1] 1 read 1 write
a[1] = a[0] 1 read 1 write
a[0] = t 1 write
----------------
7 read 5 write
If a compiler re-read data with normal insertion sort it would be
9 read 5 write
In which case the binary search would save some time.
The expected answer to this question is b), but the explanation is not precise enough:
locating the position where to insert the j-th element indeed requires log(j) comparisons instead of j comparisons for regular Insertion Sort.
inserting the elements requires j element moves in the worst case for both implementations (reverse sorted array).
Summing these over the whole array produces:
n log(n) comparisons for this modified Insertion Sort idea in all cases vs: n2 comparisons in the worst case (already sorted array) for the classic implementation.
n2 element moves in the worst case in both implementations (reverse sorted array).
note that in the classic implementation the sum of the number of comparisons and element moves is constant.
Merge Sort on the other hand uses approximately n log(n) comparisons and n log(n) element moves in all cases.
Therefore the claim the resulting insertion sort is asymptotically as good as mergesort in the worst case scenario is False, indeed because the modified Insertion Sort method still performs n2 element moves in the worst case, which is asymptotically much worse than n log(n) moves.
Note however that depending on the relative cost of comparisons and element moves, the performance of this modified Insertion Sort approach may be much better than the classic implementation, for example sorting an array of string pointers containing URLs to the same site, the cost of comparing strings with a long initial substring is much greater than moving a single pointer.
Given an integer n such that (1<=n<=10^18)
We need to calculate f(1)+f(2)+f(3)+f(4)+....+f(n).
f(x) is given as :-
Say, x = 1112222333,
then f(x)=1002000300.
Whenever we see a contiguous subsequence of same numbers, we replace it with the first number and zeroes all behind it.
Formally, f(x) = Sum over all (first element of the contiguous subsequence * 10^i ), where i is the index of first element from left of a particular contiguous subsequence.
f(x)=1*10^9 + 2*10^6 + 3*10^2 = 1002000300.
In, x=1112222333,
Element at index '9':-1
and so on...
We follow zero based indexing :-)
For, x=1234.
Element at index-'0':-4,element at index -'1':3,element at index '2':-2,element at index 3:-1
How to calculate f(1)+f(2)+f(3)+....+f(n)?
I want to generate an algorithm which calculates this sum efficiently.
There is nothing to calculate.
Multiplying each position in the array od numbers will yeild thebsame number.
So all you want to do is end up with 0s on a repeated number
IE lets populate some static values in an array in psuedo code
$As[1]='0'
$As[2]='00'
$As[3]='000'
...etc
$As[18]='000000000000000000'```
these are the "results" of 10^index
Given a value n of `1234`
```1&000 + 2&00 +3 & 0 + 4```
Results in `1234`
So, if you are putting this on a chip, then probably your most efficient method is to do a bitwise XOR between each register and the next up the line as a single operation
Then you will have 0s in all the spots you care about, and just retrive the values in the registers with a 1
In code, I think it would be most efficient to do the following
```$n = arbitrary value 11223334
$x=$n*10
$zeros=($x-$n)/10```
Okay yeah we can just do bit shifting to get a value like 100200300400 etc.
To approach this problem, it could help to begin with one digit numbers and see what sum you get.
I mean like this:
Let's say, we define , then we have:
F(1)= 45 # =10*9/2 by Euler's sum formula
F(2)= F(1)*9 + F(1)*100 # F(1)*9 is the part that comes from the last digit
# because for each of the 10 possible digits in the
# first position, we have 9 digits in the last
# because both can't be equal and so one out of ten
# becomse zero. F(1)*100 comes from the leading digit
# which is multiplied by 100 (10 because we add the
# second digit and another factor of 10 because we
# get the digit ten times in that position)
If you now continue with this scheme, for k>=1 in general you get
F(k+1)= F(k)*100+10^(k-1)*45*9
The rest is probably straightforward.
Can you tell me, which Hackerrank task this is? I guess one of the Project Euler tasks right?
I am trying to rank these functions — 2n, n100, (n + 1)2, n·lg(n), 100n, n!, lg(n), and n99 + n98 — so that each function is the big-O of the next function, but I do not know a method of determining if one function is the big-O of another. I'd really appreciate if someone could explain how I would go about doing this.
Assuming you have some programming background. Say you have below code:
void SomeMethod(int x)
{
for(int i = 0; i< x; i++)
{
// Do Some Work
}
}
Notice that the loop runs for x iterations. Generalizing, we say that you will get the solution after N iterations (where N will be the value of x ex: number of items in array/input etc).
so This type of implementation/algorithm is said to have Time Complexity of Order of N written as O(n)
Similarly, a Nested For (2 Loops) is O(n-squared) => O(n^2)
If you have Binary decisions made and you reduce possibilities into halves and pick only one half for solution. Then complexity is O(log n)
Found this link to be interesting.
For: Himanshu
While the Link explains how log(base2)N complexity comes into picture very well, Lets me put the same in my words.
Suppose you have a Pre-Sorted List like:
1,2,3,4,5,6,7,8,9,10
Now, you have been asked to Find whether 10 exists in the list. The first solution that comes to mind is Loop through the list and Find it. Which means O(n). Can it be made better?
Approach 1:
As we know that List of already sorted in ascending order So:
Break list at center (say at 5).
Compare the value of Center (5) with the Search Value (10).
If Center Value == Search Value => Item Found
If Center < Search Value => Do above steps for Right Half of the List
If Center > Search Value => Do above steps for Left Half of the List
For this simple example we will find 10 after doing 3 or 4 breaks (at: 5 then 8 then 9) (depending on how you implement)
That means For N = 10 Items - Search time was 3 (or 4). Putting some mathematics over here;
2^3 + 2 = 10 for simplicity sake lets say
2^3 = 10 (nearly equals --- this is just to do simple Logarithms base 2)
This can be re-written as:
Log-Base-2 10 = 3 (again nearly)
We know 10 was number of items & 3 was the number of breaks/lookup we had to do to find item. It Becomes
log N = K
That is the Complexity of the alogorithm above. O(log N)
Generally when a loop is nested we multiply the values as O(outerloop max value * innerloop max value) n so on. egfor (i to n){ for(j to k){}} here meaning if youll say for i=1 j=1 to k i.e. 1 * k next i=2,j=1 to k so i.e. the O(max(i)*max(j)) implies O(n*k).. Further, if you want to find order you need to recall basic operations with logarithmic usage like O(n+n(addition)) <O(n*n(multiplication)) for log it minimizes the value in it saying O(log n) <O(n) <O(n+n(addition)) <O(n*n(multiplication)) and so on. By this way you can acheive with other functions as well.
Approach should be better first generalised the equation for calculating time complexity. liken! =n*(n-1)*(n-2)*..n-(n-1)so somewhere O(nk) would be generalised formated worst case complexity like this way you can compare if k=2 then O(nk) =O(n*n)
Consider the following pseudo code:
a <- [0,0,0] (initializing a 3d vector to zeros)
b <- [0,0,0] (initializing a 3d vector to zeros)
c <- a . b (Dot product of two vectors)
In the above pseudo code, what is the flop count (i.e. number floating point operations)?
More generally, what I want to know is whether initialization of variables counts towards the total floating point operations or not, when looking at an algorithm's complexity.
In your case, both a and b vectors are zeros and I don't think that it is a good idea to use zeros to describe or explain the flops operation.
I would say that given vector a with entries a1,a2 and a3, and also given vector b with entries b1, b2, b3. The dot product of the two vectors is equal to aTb that gives
aTb = a1*b1+a2*b2+a3*b3
Here we have 3 multiplication operations
(i.e: a1*b1, a2*b2, a3*b3) and 2 addition operations. In total we have 5 operations or 5 flops.
If we want to generalize this example for n dimensional vectors a_n and b_n, we would have n times multiplication operations and n-1 times addition operations. In total we would end up with n+n-1 = 2n-1 operations or flops.
I hope the example I used above gives you the intuition.
I need a simple way to randomly select a letter from the alphabet, weighted on the percentage I want it to come up. For example, I want the letter 'E' to come up in the random function 5.9% of the time, but I only want 'Z' to come up 0.3% of the time (and so on, based on the average occurrence of each letter in the alphabet). Any suggestions? The only way I see is to populate an array with, say, 10000 letters (590 'E's, 3 'Z's, and so on) and then randomly select an letter from that array, but it seems memory intensive and clumsy.
Not sure if this would work, but it seems like it might do the trick:
Take your list of letters and frequencies and sort them from
smallest frequency to largest.
Create a 26 element array where each element n contains the sum of all previous weights and the element n from the list of frequencies. Make note of the sum in the
last element of the array
Generate a random number between 0 and the sum you made note of above
Do a binary search of the array of sums until you reach the element where that number would fall
That's a little hard to follow, so it would be something like this:
if you have a 5 letter alphabet with these frequencies, a = 5%, b = 20%, c = 10%, d = 40%, e = 25%, sort them by frequency: a,c,b,e,d
Keep a running sum of the elements: 5, 15, 35, 60, 100
Generate a random number between 0 and 100. Say it came out 22.
Do a binary search for the element where 22 would fall. In this case it would be between element 2 and 3, which would be the letter "b" (rounding up is what you want here, I think)
You've already acknowledged the tradeoff between space and speed, so I won't get into that.
If you can calculate the frequency of each letter a priori, then you can pre-generate an array (or dynamically create and fill an array once) to scale up with your desired level of precision.
Since you used percentages with a single digit of precision after the decimal point, then consider an array of 1000 entries. Each index represents one tenth of one percent of frequency. So you'd have letter[0] to letter[82] equal to 'a', letter[83] to letter[97] equal to 'b', and so on up until letter[999] equal to 'z'. (Values according to Relative frequencies of letters in the English language)
Now generate a random number between 0 and 1 (using whatever favourite PRNG you have, assuming uniform distribution) and multiply the result by 1000. That gives you the index into your array, and your weighted-random letter.
Use the method explained here. Alas this is for Python but could be rewritten for C etc.
https://stackoverflow.com/a/4113400/129202
First you need to make a NSDicationary of the letters and their frequencies;
I'll explain it with an example:
let's say your dictionary is something like this:
{#"a": #0.2, #"b", #0.5, #"c": #0.3};
So the frequency of you letters covers the interval of [0, 1] this way:
a->[0, 0.2] + b->[0.2, 0.7] + c->[0.7, 1]
You generate a random number between 0 and 1. Then easily by checking that this random belongs to which interval and returning the corresponding letter you get what you want.
you seed the random function at the beginning of you program: srand48(time(0));
-(NSSting *)weightedRandomForDicLetters:(NSDictionary *)letterFreq
{
double randomNumber = drand48();
double endOfInterval = 0;
for (NSString *letter in dic){
endOfInterval += [[letterFreq objectForKey:letter] doubleValue];
if (randomNumber < endOfInterval) {
return letter;
}
}
}