Questions tagged [GPGPU] (113)

4
answers

Is it possible to use multiple graphics cards for acceleration of scientific computing in the languages C++ or Python?

At the moment I'm doing scientific calculations (think of the matrix, which is good parallelise). If initially all began with sequential calculations on the CPU in python, pretty soon came the idea of parallelization on the CPU, and then CUDA (using a library numba in python). And the results are very good. However, it got ...
enola_Konopelski74 asked April 7th 20 at 09:14
1
answer

What is best for Premiere: 4770 (4.5 GHz) or 1070?

Good afternoon. A CPU i7 4770k overclocked to 4.5 GHz, 16GB DDR3 memory 2000мгц (11-11-11-28) gtx 1070 jetstream, the power profile to max performance, if that. Editing videos in premiere, 1 video track, 1 audio, no effects, just the video sped up from 6K to 10K. Bogged down playback, delay before playing for 2-4 seconds, s...
roberta.Hue asked April 3rd 20 at 17:28
2
answers

It is advantageous if the calculation of photogrammetry graphics cards for mining?

Having tasks that need to be considered on the cards. Wanted to know if the card is sharpened for mining, no video outputs (these were from nvidia, it seems) for such things. Maybe someone was testing on similar computational processes? Or can anyone have access to such computing systems to test yourself :)
ken_Leuschke asked April 2nd 20 at 16:19
2
answers

Video card P104-100 can not be used for mining?

Now popular cards on Yandex market P104-100 1607Mhz PCI-E 1.1 4096Mb, this card is issued specifically for mining, do they still buying miners? I figured on the maps and Bang! (I'm not a miner). Now collect the server, and the programmer is asked to supply graphics accelerator with CUDA for some work with the acceleration d...
Marilou.Pouros asked April 1st 20 at 16:08
1
answer

Is it possible and does it make sense to move this code to OpenCL?

Good day! There's a piece of code:struct TTour { QVector<int> tour; double cost; }; TTour Tournamet::iter(const TTour &currentTour) { int n = currentTour.tour.count(); TTour bestTour = currentTour; for(int i=1; i<n-2; ++i){ for(int j=i+1; j<n+1; ++j) { if(j-i == 1) continue; QVector<int> swap ...
Osborne_Padbe asked March 31st 20 at 16:00
1
answer

What is a good website with the features of graphic processors?

Hello! Tell me good sites where you can see the characteristics of different models of GPUs? I found a good resource on conventional processors: www.cpu-world.com/index.html I would like something similar, only for GPUs. Share, who knows what! I think it will be a lot useful.
anabel32 asked March 24th 20 at 12:08
0
answer

How to fix error "clCreateCommandQueue is deprecated" in Visual Studio 2017?

There is a program in OpenCL 2.0. When compiling appears an error: The most popular solution is to declare at the beginning of a preprocessor macro :#define CL_USE_DEPRECATED_OPENCL_1_2_APIS #include <CL/cl.h> To me it didn't help. Later, I tried to call another one, relevant for OpenCL 2.0 feature clCreateCommandQue...
destin.Sawayn asked March 24th 20 at 11:05
0
answer

The truth is that OMP inhibits CUDA?

In General the problem is. There is a banal code**** double time = MPI_Wtime(); cudaMemcpy(*****) cout << "time =" << MPI_Wtime() - time << endl; Nvcc compile with mpicxx -fopenmp Set the variable OMP_NUM_THREADS = 1 Measured the time. Next, Set the variable OMP_NUM_THREADS = 2 Measured the time, it was w...
lillie asked March 21st 20 at 13:19
0
answer

How to connect Cuda in Qt?

Hello, recently I finished writing a neural network in Cuda for recognition of handwritten digits and now the question arose about the graphic shell program. While just started to learn Qt and was immediately faced with the fact that you don't know how to connect Cuda in Qt. Just one request, please explain everything in gr...
Zack_Witting asked March 20th 20 at 12:52
0
answer

Error when calling a Cuda kernel?

Good afternoon, in the program class:class neuralNet { float input_nodes; float hidden_nodes; float output_nodes; float learning_grade; float* wih; float* who; public: neuralNet(int in, int hid, int out, float lr); ~neuralNet(); float*activation_function(float* a, int n); void train(float*inputs, float*targets); ...
Berniece_Cron asked March 20th 20 at 11:23