C-Programming | C Programming Experiments | GPU library

The CUDA programming guide gives non-device-specific maxima for grid dimensions in blocks: 2^31 - 1 for the x dimension, 2^16 - 1 for the y and z dimensions (table 15, as of CUDA 11.3).

My question: Where are these values defined, in code? I also looked at the driver API entry on cudaLaunchKernel, and it doesn't mention such constants either. I searched for 65535, for "<< 16" and for "<<16" in the CUDA header files, and no luck there either.

I was trying to solve the problem Modulo strength at hackerearth ,
https://www.hackerearth.com/practice/basic-programming/implementation/basics-of-implementation/practice-problems/golf/modulo-strength-4/ ,
so basically we have to find all such pairs of no. (say i,j) such that A[i]%k=A[j]%k where k is a no. given in the question ,
i tried brute force approach and got time limit exceeded at some of the last test cases and
in the discussion tab i found a code which is working but i couldn't understand what exactly it does, and the underlying thinking behind the algorithm used.

Rules for variable names in C are as follows:

1. A variable name can only have letters (both uppercase and lowercase letters), digits and underscore.
2. The first letter of a variable should be either a letter or an underscore.
3. There is no rule on how long a variable name (identifier) can be. However, you may run into problems in some compilers if the variable name is longer than 31 characters. (source: [https://www.programiz.com/c-programming/c-variables-constants])

I'm wondering about whether, theoretically, if a single underbar(_) or double underbar(__) be used as a variable? and can printf or scanf be used as a variable?

While playing with the c compiler(Dev C++) and linux Ubuntu Vi, even if I used the above as a variable name, there weren't any errors or warnings.

The code I used is as follows:

I have two graph classes: DirectedGraph and DirectedBreakableGraph. DirectedBreakableGraph inherits from DirectedGraph, and provides the ability to temporarily break edges, meaning that they should not be traversed, even though they are still part of the graph. This way cycles in the graph can temporarily be resolved.

DirectedGraph has a DFS (depth-first search) method for traversing the graph. This method is virtual, so that DirectedBreakableGraph can have its own implementation, ignoring broken edges.

I want to be able to specify from which node(s) the depth-first search should start, by passing a collection of start nodes to DFS. Now, I do not want to put restrictions on the type of the collection containing the start nodes. It could for example be a QVector or a QSet.

Since the method is virtual, it cannot also be a template. Is it then impossible to let it accept both QVector and QSet as input parameter? Maybe apart from using type erasure?

I am learning about pointers and how they work in general. I have found some example code online: https://overiq.com/c-programming-101/pointers-and-2-d-arrays/

I've been trying to get some of my other code to run, and I've run into a bit of an issue with dynamic shared memory. According to documentation (https://docs.nvidia.com/cuda/cuda-c-programming-guide/index.html#shared), I'm supposed to allocate one array of memory, and then typecast pointers to specific locations in that array like this:

I was learning and practising c-programming and I got an idea to print a player in a battle field made using 'x' and '0' the '0' is the place where the player 'x' can be placed and I want to be able to place multiple 'x' player in one battle field using variable argument in c. I thought to create this to exercise how I could simplify my code if I needed too. When I insert only one 'x' its fine but when I insert multiple 'x' this code doesn't work.(I mean I can create structure and pass it and put as many 'x' as I want to but this way is the lengthy way) So I used variable argument. Any idea on how I could modify this code to put two or more 'x' in same battle field in the easiest way possible. I was also thinking of using file handling, putting the positionx and positiony values in txt file and using it.

Can this work using variable argument? If yes how? If no then how should I do it in very effective and fast way? My code is (my code looks like it is working but the position of 'man2' is not right :

So I tried to follow an example from: Fixed Point Arithmetic in C Programming, but I get the following error left shift count >= width of type.

I did find others having this issue here on stackoverflow but I don't understand this error?

Follow-Up question to PyTorch: Dynamic Programming as Tensor Operation.

Could the following be written as a tensor operation instead of a loop?