51
edits
Changes
→Sources
[[File:latestGpuSpecsAmd.jpg|900px|]]
[[File:latestGpuSpecsNvidia2.jpg|900px|]]
'''RX 6900 XT vs RTX 3090: Specifications''':
[[File:latestGpuSpecs.jpg]]
== Means of parallelisation on GPUs ==
https://stackoverflow.com/questions/7263193/opencl-vs-openmp-performance#7263823
== Instructions for NVIDEA ==
'''How to set up the compiler and target offloading for Linux with a target NVIDIA GPU'''
$ clang -fopenmp -fopenmp-targets=nvptx64 -O2 foo.c
</pre>
== Instructions for NVIDEA ==
'''How to set up compiler and target offloading for Linux on AMD GPU: (Elena)'''
AOMP is an open source Clang/LLVM based compiler with added support for the OpenMP® API on Radeon™ GPUs. Use this repository for releases, issues, documentation, packaging, and examples,.
== Programming GPUs with OpenMP ==
<h3>Teams construct</h3>
In order to provide parallelization within the GPU architectures there is an additional construct known as the ''teams'' construct, which creates multiple master threads on the device. [[File: Teams.JPG|thumb|upright=1.2|right|alt=OpenMP teams]]
Each master thread can spawn a team of its own threads within a parallel region. But threads from different teams cannot synchronize with other threads outside of their own team.
[[File: Distribute.JPG|thumb|upright=1.2|right|alt=OpenMP distribute]]
<pre>
int main() {
</pre>
== Results and Graphs (Nathan/Elena) ==
https://www.pcmag.com/encyclopedia/term/core-i7 AMD RX-580 GPU architecture
https://premiumbuilds.com/comparisons/rx-6900-xt-vs-rtx-3090/ -> compare Flagship GPU's 2020
[http://www.nvidia.com/en-us/geforce/graphics-cards/30-series/ http://www.nvidia.com/en-us/geforce/graphics-cards/30-series/ nvidia]
[https://www.pcmag.com/encyclopedia/term/core-i7 https://www.pcmag.com/encyclopedia/term/core-i7 CPU picture ]
https://rocmdocs.amd.com/en/latest/Programming_Guides/Programming-Guides.html?highlight=hip <- HIP, openCl
