How do I generate missing values of an array in Fortran? - fortran

I have an array, x, with dimension 4, that has the following values
3, 4.5, 7, 9
How do I generate the missing values, and make a new one like this?
3, 3.5, 4, 4.5, 5, 5.5, ...,9
Thanks in advance.

Related

How to put in variable as list

I'm trying to make a program that for a sublist of numbers, uses index as a variable and selects each number from the list of lists
so if my numbest = [[1, 2, 3, 4, 5], [2, 4, 6, 8, 10], [3, 5, 7, 9, 11]]
I want to be able to call the function like this
column_sum(2, [[1, 2, 3, 4, 5], [2, 4, 6, 8, 10], [3, 5, 7, 9, 11]]) will add the numbers at index 2 in each sublist (3, 6, and 7) and will return the number 16."
I can't for the life of me figure out how to print
for i in numlist:
print numbest[index]
Looks like Python, so imma say that all you need to do is have a variable that is a running total, add up all the numbers that are the values at the index you specify, and then return that value.
Alexander is also right and if his way is easier for you, you can find resources https://www.w3schools.com/python/ref_func_sum.asp and https://www.w3schools.com/python/python_lists_comprehension.asp

How to get a index of element in the array with Dlang?

I would like to get a index of array satisfying the condition.
Then, I'd like to get max of them.
With Ruby:
# normal array
array.index{|n| n>W }
# 2-dimensional array
matrix.map{|arr| arr.index{|n| n>W}}
How to do this with Dlang ?
You can use countUntil, it accepts a predicate :
[1, 2, 3, 4, 5, 6, 7, 8].countUntil!(c => c > 5).writeln;
To get the index of the max element, use the aptly name maxIndex function :
[1, 2, 3, 4, 5, 6, 7, 8].maxIndex.writeln;

Memory access error with _mm512_i64gather_pd()

I am trying to use a very simple example of the AVX-512 gather instructions:
double __attribute__((aligned(64))) array3[17] = {1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0,
9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0,
17.0};
int __attribute__((aligned(64))) i_index_ar[16] = {1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16};
__m512i i_index = _mm512_load_epi64(i_index_ar);
__m512d a7AVX = _mm512_i64gather_pd(i_index, &array3[0], 1);
Unfortunetly, my last call to _mm512_i64gather_pd results in an memory access error (memory dumped).
Error message in German: Speicherzugriffsfehler (Speicherabzug geschrieben)
I am using Intel Xeon Phi (KNL) 7210.
edit: The error here was, that I was using 32 bit integers with 64bit load instructions and scale in _mm512_i64gather_pd has to be 8 or sizeof(double).
I think you need to set scale to sizeof(double), not 1.
Change:
__m512d a7AVX = _mm512_i64gather_pd(i_index, &array3[0], 1);
to:
__m512d a7AVX = _mm512_i64gather_pd(i_index, &array3[0], sizeof(double));
See also: this question and its answers for a fuller explanation of Intel SIMD gathered loads and their usage.
—
Another problem: your indices need to be 64 bit ints, so change:
int __attribute__((aligned(64))) i_index_ar[16] = {1, 2, 3, 4, 5, 6, 7, 8, 9, ...
to:
int64_t __attribute__((aligned(64))) i_index_ar[16] = {1, 2, 3, 4, 5, 6, 7, 8, 9, ...

Efficient Eigen Matrix SubIndexing + Concatenation

I'm using Eigen for easy optimization of some of my matrix math. I'm currently trying to make the following operation more efficient:
Given Matrix A:
1, 2, 3
4, 5, 6
Matrix B:
7, 11, 13, 19, 26, 7, 11
8, 9, 15, 6, 8, 4, 1
and "index map" column vector IM:
0, 1, 3, 6
I'd like to append the columns of Matrix B mapping to the indexes in IM, to Matrix A as such:
1, 2, 3, 7, 11, 19, 11
4, 5, 6, 8, 9, 6, 1
I'm currently able to do this with a massive for loop, but this is the bottleneck in my code and I'd like to avoid this:
#pragma unroll
for (int i = 0; i < 25088; i++) {
block.noalias() += _features.col(ff[i]);
}
I've seen the discussion here and poured over the docs but can't seem to figure out the right syntax relating to Eigen matrices: http://eigen.tuxfamily.org/bz/show_bug.cgi?id=329
Any thoughts/tips would be much appreciated!

Trying to simulate python combinations in C++ with next_permutation

I need to port a snippet written in Python to C++
but that snippet is using combinations from itertools in python.
The line that I'm really interested to porting over to C++ is this one:
for k in combinations(range(n-i),2*i):
range(n-i) in Python will generate a list from 0 to (n-i) - 1
Let n = 16, i = 5
print range(n-i)
outputs:
[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10]
and python combinations will generate all possible combinations in that list.
e.g.
print list(combinations(range(n-i),2*i))
outputs:
[(0, 1, 2, 3, 4, 5, 6, 7, 8, 9),
(0, 1, 2, 3, 4, 5, 6, 7, 8, 10),
(0, 1, 2, 3, 4, 5, 6, 7, 9, 10),
(0, 1, 2, 3, 4, 5, 6, 8, 9, 10),
(0, 1, 2, 3, 4, 5, 7, 8, 9, 10),
(0, 1, 2, 3, 4, 6, 7, 8, 9, 10),
(0, 1, 2, 3, 5, 6, 7, 8, 9, 10),
(0, 1, 2, 4, 5, 6, 7, 8, 9, 10),
(0, 1, 3, 4, 5, 6, 7, 8, 9, 10),
(0, 2, 3, 4, 5, 6, 7, 8, 9, 10),
(1, 2, 3, 4, 5, 6, 7, 8, 9, 10)]
I want to generate similar output using std::vector and next_permutation from C++ but I'm still getting erroneous results. This is my current approach:
for(int j = 0; j < n-i; j++) {
temp_vector.push_back(j);
}
That snippet is equivalent to range(n-i) in Python.
But the following snippet:
do {
myvector.push_back(temp_vector);
} while(next_permutation(temp_vector.begin(),temp_vector.begin()+2*i));
cout<<myvector.size()<<endl;
Is not equivalent to combinations(range(n-i),2*i)) in Python, and I've tried many variations and still haven't been able to come up with the results I'm expecting.
For example:
Let n = 16
i = 5
Python
>>> print len(list(combinations(range(n-i),2*i)))
11
C++
#include <vector>
#include <iostream>
using namespace std;
int main() {
vector<int> temp_vector;
vector< vector<int> > myvector;
int n = 16, i = 5;
for(int j = 0; j < n - i; j++) {
temp_vector.push_back(j);
}
do {
myvector.push_back(temp_vector);
} while(next_permutation(temp_vector.begin(), temp_vector.begin()+2*i));
cout<<myvector.size()<<endl;
return 0;
}
g++ combinations.cpp
./a.out
3628800
Any guidance will be greatly appreciated! Thanks a lot!
combinations and permutations are not the same thing.
A combination is an unordered list of a subset of the items from another set. A permutation is a unique order of the items in the list.
You're generating all combinations of 10 things from a list of 11 things, so you'll get 11 results, each one missing a different one of the original 11 items.
Generating every permutation will generate every unique order of the original 11 items. Since the items in this case are all unique that means the result would be 11! lists where each contains all 11 items. You're only generating permutations from the first 10 items however, so you're getting 10! lists, none of which contain the 11th item.
You need to find an algorithm for generating combinations instead of permutations.
There's no built-in algorithm for combinations. std::next_permutation can be used as part of an algorithm to generate combinations: See Generating combinations in c++.
Here's an old draft proposal for algorithms for combinations, including code.