Applied Parallel Computing LLC

Can you imagine interactive Jupyter notebooks just inside your static weblog without any server-side? Now possible with Iodide by Mozilla! Here are some easy steps to put it into your GitHub Pages.

Embedding Jupyter-quaility rich visual Python into a static website We all love our CV/blog websites hosted on GitHub Pages. We also love Jupyter notebooks for revolting the look and feel of daily data…

Results of scientific or engineering simulations are often needed to be recorded on-the-fly. That is, the volume of data is often inapplicable to saving all individual frames on disk and postprocess them into video afterwards. This library leverages FFmpeg to embed video encoding directly into frame-producing application.

apc-llc/moviemaker-cpp moviemaker-cpp - Encode video frame by frame in C++ using FFmpeg

Recent 5.x and 6.x GCC compilers are causing NVCC to produce weird compile errors related to avx512 intrinsics headers. Here we show how to quickly fix the issue.

How to fix CUDA and avx512vlintrin.h incompatibilty issue CUDA Trainings

Did you know that we actually can leverage Beignet in order to run OpenCL on older Intel GPUs, like Ivy Bridge '2012? Read more in great GitHub post by Piotr Pomienski:

How-to: OpenCL 1.2 on Intel HD Graphics (3. gen and newer) under Linux · Issue #53 · HPCE/hpce-2017-cw3 Warning: a lot of text. TL;DR: Intel only provides OpenCL for Windows, there's an open source alternative for Linux but it doesn't always work due to OpenCL 2.0 being used by default, so co...

https://parallel-computing.pro/index.php/12-hardware/66-test-3-ai-toolkits-for-free

Test 3 AI toolkits for free CUDA Trainings

Applied Parallel Computing LLC offers a specialized 4-day course on GPU-enabled Neural Networks. The course is intended for developers willing to rapidly get NVIDIA-based AI technology into their new and existing software solutions.

New 4-day Training Course on GPU-enabled Neural Networks CUDA Trainings

GPU-equipped clusters are often managed by SLURM. In order to debug/profile an application, developer is allowed to SSH from the frontend node to compute nodes granted for ex*****on. How this pipeline can play together with NVIDIA Visual Profiler?

Remote profiling with NVIDIA Visual Profiler on a SLURM-based cluster

Using CUDA device functions from OpenACC Due to high-level nature, OpenACC does not offer access to useful GPU-specific features. In this article we demonstrate how to call CUDA device functions from within OpenACC kernels by two examples: GPU compute grid retrieval and printf.

CUDA-like runtime interface for Xeon Phi Unlike for GPUs, Intel has left OpenMP offloading directives as the only Xeon Phi programming option. Based on liboffloadmic, we have prototyped "micrt" - a programming interface to execute memory transfers and kernels on Xeon Phi, similarly to CUDA.