GPGPU | Small library for running arbitrary computations | GPU library

Instagram @sennepldn

Fluffy ball demo (from Instagram)

Fluffy ball demo (If you can't access Instagram Link)

When I saw this example three days ago, I was very surprised, so I wanted to challenge myself and try to make this effect.

After three days, I have tried many ways like using vertex shader or GPUComputationRenderer, but still did not make the correct simulation. So I finally decided to come up and ask you all.

First, use the following two balls that seem to have hard spikes to simulate the Fluffy Ball in the Demo.

From the first frame to the second frame, the ball rotated about 30 degrees clockwise (I randomly assumed an angle)

Simplify the upper ball into the picture below.

The black line in the picture represents the longest spike in the picture above.

A1, a2, b1, b2 (Vector3) represent the positions of points a and b at different times.

However, in the final simulation, this is a soft hair instead of a hard thorn, so the position of Frame2 b point should be b3.

So I plan to use the following method to simulate.

FAKE CODE (in vertex shader) (b1, b2, b3 means position in vec3)

I'm writing this simple cuda code and I'm unable to compile it. The code contains part of code written in C. This is the structure of the program:

read_data.c file contains a function called read_data add.cu file contains a function called add (this is the part that should run in the GPGPU) optimize.h file contains the necessary headers. master.c file contains the main function.

The optimize.h file looks like follows:

I am attempting to use use WebGL2 for some GPGPU computations. A key component of this is setting the value of a texel to the bitwise-OR of itself and the new value computed in the fragment shader. Is there a way of applying this bitwise operation to each fragment instead of overwriting the value completely?

Here is the relevant code:

I am trying to output more than one buffer from a shader - the general goal is to use it for GPGPU purposes. I've looked at this answer and got closer to the goal with this:

Background
I've experimented with OpenCL with the C++ programming language on a Windows PC to write simple programs to a PC's GPU. In other words, I used it as a GPGPU. Under simple calculations I mean that I made two arrays, each containing 1 million items, than I added the corresponding parts of the two arrays.
So if I have X[1000000] and Y[1000000], then I would do:

I am preparing to buy a cluster of SOPINE A64 modules for basic (CPU-based) parallel computing, and I noticed that the modules also have GPUs. It wasn't difficult to find that the Mali-400 is not compatible with OpenCL, but I'm having trouble confirming that I would be able to use the OpenGL interface for general-purpose GPU programming. I don't need to do anything fancy; I just want to know if I can offload some of my matrix-heavy tasks to the GPU.

I can find a helpful tutorial on GPGPU programming in OpenGL but it assumes access to GLUT, which isn't available on OpenGL ES 2.0, and the most relevant answer I've found on SE talks about doing what I want on iOS but not with the same GPU.

Is it as simple as using something besides GLUT to set up the OpenGL environment and then following the linked tutorial? Or are there other hardware limitations I need to be aware of?

I am new to GPGPU, and I have a confusion about SAXPY function: The SAXPY function is like given two same sizes and type vector X and Y, do the operation that changes each element in Y: y[i] = y[i]+a*x[i]

I am not sure of can we change SAXPY's formula, like: y[i]=(y[i]+a)*(x[i]+c)

but in this case, there is a new constant c, I have no idea how to call SAXPY in this condition.

Thanks for your valuable time.

I'm using cupy with Spyder3.3.6 and Python 3.7.5 in Windows machine(Win10 Pro 64bit, i7-7700, 8GBMemory, GTX-1060-6GB).
Version of cupy, chainer, cuda and cuDNN are 6.0.0, 5.3.0, 10.1.243, and 7.6.4,respectively.

When I import cupy, this error has occured:

I am looking for a standardized approach to stream JPG images over the network. Also desirable would be a C++ programming interface, which can be easily integrated into existing software.

I am working on a GPGPU program which processes digitized signals and compresses them to JPG images. The size of the images can be defined by a user, typically the images are 1024 x 2048 or 2048 x 4096 pixel. I have written my "own" protocol, which first sends a header (image size in bytes, width, height and corresponding channel) and then the JPG data itself via TCP. After that the receiver sends a confirmation that all data are received and displayed correctly, so that the next image can be sent. So far so good, unfortunately my approach reaches just 12 fps, which does not satisfy the project requirements.

I am sure that there are better approaches that have higher frame rates. Which approach do streaming services like Netflix and Amzon take for UHD videos? Of course I googled a lot, but I couldn't find any satisfactory results.