I want to shuffle a 3d array on the third dimension using Shuffle.c.
Till now I have used Shuffle(arr,3) with great performance. Now I try to do the same, but with array of Complex numbers and get this Error:
*** Shuffle[mex]: Use index mode for complex input!
I haven't found the proper way to use index mode.
Thank you.
Did you verify the better performance for you input sizes? For 3D matrices shuffling one dimensions, you are better off not using the MEX function with the above mentioned syntax. It forces MATLAB to copy the whole matrix to mex and back. For comparison:
arr=rand(100,100,100);
t.mexshuffle=timeit(#()(Shuffle(arr,3)));
t.randperm=timeit(#()arr(:,:,randperm(size(arr,3))));
t.mexindexshuffle=timeit(#()arr(:,:,Shuffle(size(arr,3),'index')));
The results are:
struct with fields:
mexshuffle: 0.0183
randperm: 0.0038
mexindexshuffle: 0.0037
It does not really matter if you use Shuffle with the index option or randperm, but directly using Shuffle is slower. Nice side-effect, the later two options support complex numbers.
Above code might be a bit hard to read, here a cleaner version of the suggested solution:
P=randperm(size(arr,3)); % Permutation vector, use whichever generator you prefer.
%P=Shuffle(size(arr,3),'index');
arr_out=arr(:,:,P);
Related
I would like to optimize this piece of Matlab code but so far I have failed. I have tried different combinations of repmat and sums and cumsums, but all my attempts seem to not give the correct result. I would appreciate some expert guidance on this tough problem.
S=1000; T=10;
X=rand(T,S),
X=sort(X,1,'ascend');
Result=zeros(S,1);
for c=1:T-1
for cc=c+1:T
d=(X(cc,:)-X(c,:))-(cc-c)/T;
Result=Result+abs(d');
end
end
Basically I create 1000 vectors of 10 random numbers, and for each vector I calculate for each pair of values (say the mth and the nth) the difference between them, minus the difference (n-m). I sum over of possible pairs and I return the result for every vector.
I hope this explanation is clear,
Thanks a lot in advance.
It is at least easy to vectorize your inner loop:
Result=zeros(S,1);
for c=1:T-1
d=(X(c+1:T,:)-X(c,:))-((c+1:T)'-c)./T;
Result=Result+sum(abs(d),1)';
end
Here, I'm using the new automatic singleton expansion. If you have an older version of MATLAB you'll need to use bsxfun for two of the subtraction operations. For example, X(c+1:T,:)-X(c,:) is the same as bsxfun(#minus,X(c+1:T,:),X(c,:)).
What is happening in the bit of code is that instead of looping cc=c+1:T, we take all of those indices at once. So I simply replaced cc for c+1:T. d is then a matrix with multiple rows (9 in the first iteration, and one fewer in each subsequent iteration).
Surprisingly, this is slower than the double loop, and similar in speed to Jodag's answer.
Next, we can try to improve indexing. Note that the code above extracts data row-wise from the matrix. MATLAB stores data column-wise. So it's more efficient to extract a column than a row from a matrix. Let's transpose X:
X=X';
Result=zeros(S,1);
for c=1:T-1
d=(X(:,c+1:T)-X(:,c))-((c+1:T)-c)./T;
Result=Result+sum(abs(d),2);
end
This is more than twice as fast as the code that indexes row-wise.
But of course the same trick can be applied to the code in the question, speeding it up by about 50%:
X=X';
Result=zeros(S,1);
for c=1:T-1
for cc=c+1:T
d=(X(:,cc)-X(:,c))-(cc-c)/T;
Result=Result+abs(d);
end
end
My takeaway message from this exercise is that MATLAB's JIT compiler has improved things a lot. Back in the day any sort of loop would halt code to a grind. Today it's not necessarily the worst approach, especially if all you do is use built-in functions.
The nchoosek(v,k) function generates all combinations of the elements in v taken k at a time. We can use this to generate all possible pairs of indicies then use this to vectorize the loops. It appears that in this case the vectorization doesn't actually improve performance (at least on my machine with 2017a). Maybe someone will come up with a more efficient approach.
idx = nchoosek(1:T,2);
d = bsxfun(#minus,(X(idx(:,2),:) - X(idx(:,1),:)), (idx(:,2)-idx(:,1))/T);
Result = sum(abs(d),1)';
Update: here are the results for the running times for the different proposals (10^5 trials):
So it looks like the transformation of the matrix is the most efficient intervention, and my original double-loop implementation is, amazingly, the best compared to the vectorized versions. However, in my hands (2017a) the improvement is only 16.6% compared to the original using the mean (18.2% using the median).
Maybe there is still room for improvement?
I have a very big sparse csc_matrix x. I want to do elementwise exp() on it. Basically what I want is to get the same result as I would have got with numpy.exp(x.toarray()). But I can't do that(my memory won't allow me to convert the sparse matrix into an array). Is there any way out? Thanks in advance!
If you don't have the memory to hold x.toarray(), you don't have the memory to hold the output you're asking for. The output won't be sparse; in fact, unless your input has negative infinities in it, the output probably won't have a single 0.
It'd probably be better to compute exp(x)-1, which is as simple as
x.expm1()
If you want to do something on nonzeros only: the data attribute is writable at least in some representations including csr and csc. Some representations allow for duplicate entries, so make sure you are acting on a "normalised" form.
To change non-zero elements, maybe this would work for you:
x = some big sparse matrix
np.exp( x.data, out=x.data ) # ask np.exp() to store results in existing x.data
presumably slower:
# above seems more efficient (no new memory alloc).
x.data = np.exp( x.data )
I've been wrestling with how to get an element-wise log2() of each non-zero array element. I ended up doing smth like:
np.log2( x.data, out=x.data )
The following two techniques seem like exactly what I was looking for. My matrix is sparse but it still plenty of non-zero elements.
Credit to #DSM here for the idea of directly changing x.data, I think that is a superb insight about sparse matrices.
Credit to #Mike Müller for the idea of using "out" as itself. In the same thread, #kmario23 points out an important caveat about promoting .data to floats (inputs could be int or smth) so it is compatible with the .exp() or whatever function, I would want to do that if I was writing smth for general use.
note: I'm just starting to learn about sparse matrices, so would like to know if this is a bad idea for reason(s) I'm not seeing. Please do let me know if I'm on thin ice with this.
Normally I wouldn't mess with private attributes, but .data shows up pretty clearly in the attributes documentation for the various sparse matrices I've looked at.
I need to (repeatedly) build a vector of length 200 from a vector of length 2500. I can describe this operation using multiplication by a matrix which is extremely sparse: it is 200x2500 and has only one entry in each row. But I have very little control over where this entry is. My actual problem is that I need to apply this matrix not to the vector that I currently have, but rather to some componentwise function of this vector. Since I have all this sparsity, it is wasteful to apply this componentwise function to all 2500 components of my vector. Instead I would rather apply it only to the 200 components that actually contribute.
A program (with randomly chosen numbers replacing of my actual numbers) which would have a similar problem would be something like this:
ind=randi(2500,200,1);
coefficients=randn(200,1);
A=sparse(1:200,ind,coefficients,200,2500);
x=randn(2500,1);
y=A*subplus(x);
What I don't like here is applying subplus to all of x; I would rather only have to apply it to x(ind), since only that contributes to the matrix product.
Right now the only way I can see to work around this is to replace my sparse matrix with a 200-component vector of coefficients and a 200-component vector of indices. Working this way, the code above would become:
ind=randi(2500,200,1);
coefficients=randn(200,1);
x=randn(2500,1);
y=coefficients.*subplus(x(ind))
Is there a better way to do this, preferably one that would work when A contains a few elements per row instead of just one?
The code in your question throws an exception, I think it should be:
n=2500;
m=200;
ind=randi(n,m,1);
coefficients=randn(m,1);
A=sparse(1:m,ind,coefficients,m,n);
x=randn(n,1);
Your idea using x(ind) was basically right, but ind would reorder x which is not intended. Instead you could use sort(unique(ind)). I opted to use the sparse logical index any(A~=0) because I expect it to be faster, but you could compare both versions.
%original code
y=A*subplus(x);
.
%multiplication using sparse logical indexing:
relevant=any(A~=0);
y=A(:,relevant)*subplus(x(relevant));
.
%fixed version of your code
relevant=sort(unique(ind));
y=A(:,relevant)*subplus(x(relevant));
I am trying to vectorise a for loop. I have a set of coordinates listed in a [68x200] matrix called plt2, and I have another set of coordinates listed in a [400x1] matrix called trans1. I want to create a three dimensional array called dist1, where in dist1(:,:,1) I have all of the values of plt2 with the first value of trans1 subtracted, all the way through to the end of trans1. I have a for loop like this which works but is very slow:
for i=1:source_points;
dist1(:,:,i)=plt2-trans1(i,1);
end
Thanks for any help.
If I understood correctly, this can be easily solved with bsxfun:
dist1 = bsxfun(#minus, plt2, shiftdim(trans1,-2));
Or, if speed is important, use this equivalent version (thanks to #chappjc), which seems to be much faster:
dist1 = bsxfun(#minus, plt2, reshape(trans1,1,1,[]));
In general, bsxfun is a very useful function for cases like this. Its behaviour can be summarized as follows: for any singleton dimension of any of its two input arrays, it applies an "implicit" for loop to the other array along the same dimension. See the doc for further details.
Vectorizing is a good first optimization, and is usually much easier than going all in writing your own compiled mex-function (in c).
However, the golden middle-way for power users is Matlab Coder (this also applies to slightly harder problems than the one posted, where vectorization is more or less impossible). First, create a small m-file function around the slow code, in your case:
function dist1 = do_some_stuff(source_points,dist1,plt2,trans1)
for i=1:source_points;
dist1(:,:,i)=plt2-trans1(i,1);
end
Then create a simple wrapper function which calls do_some_stuff as well as defines the inputs. This file should really be only 5 rows, with only the bare essentials needed. Matlab Coder uses the wrapper function to understand what typical proper inputs to do_some_stuff are.
You can now fire up the Matlab Coder gui from the Apps section and simply add do_some_stuff under Entry-Point Files. Press Autodefine types and select your wrapper function. Go to build and press build, and you are good to go! This approach usually bumps up the execution speed substantially with almost no effort.
BR
Magnus
I have one big matrix of for example 3000X300. And I need to select each element and do several calculations with it. I looked into using the array fun function but because the output of my program is not one value this is not possible.
It works fine now with the loops but it has to preform much faster, so i want to remove the for loop.
Maybe i'll try to be more specific: Each value of the big matrix has to give me an answer of 4 different matrices with the size of 4X6020..
So i don't know if this is possible making this vectorized...
Maybe somebody has other suggestions to make it faster?
greetings,
You can use arrayfun and set uniformoutput to false. See here.