RadeonProRenderSDK

From Gentoo Wiki
Jump to:navigation Jump to:search

AMD Radeon ProRender is a powerful physically-based rendering engine that enables creative professionals to produce stunningly photorealistic images. https://www.amd.com/en/technologies/radeon-prorender

ProRender SDK

Repository preparation

Ebuild for SDK installation not in official gentoo repository, but can be found in feniksa overlay. Use either eselect repository or layman to configure the additional repository.

eselect repository

Install app-eselect/eselect-repository:

root #emerge --ask --noreplace app-eselect/eselect-repository

Enable and sync the feniksa repository:

root #eselect repository enable feniksa
root #emerge --sync feniksa

Layman

First, install layman:

root #emerge --ask app-portage/layman

Fetch and display a list of all the repositories available through official references:

root #layman -L

Add a feniksa repository:

root #layman -a feniksa

SDK installation

To install Radeon Pro Render SDK:

root #emerge --ask dev-libs/radeon-pro-render-sdk

Tutorials

Build

Install the package with the examples USE flag:

root #emerge --ask dev-libs/radeon-pro-render-sdk[examples]

Run

All tutorial binaries may be found at /usr/share/RadeonProRender/tutorials/Bin directory. First, let's check ProRender initialize

root #cd /usr/share/RadeonProRender/tutorials/Bin
root #./00_context_creation64D
Radeon ProRender SDK simple context creation tutorial.
RPR Context creation succeeded.
GPU0 name : NVIDIA GeForce GTX 1080

If everything is fine, last line should show GPU name. It means, that RadeonProRender correctly initialize OpenCL, found available GPU and initialized resources for PBR rendering . Source code of this tutorial may be found at /usr/share/RadeonProRender/tutorials/00_context_creation/main.cpp

Note
First run may take up to 2 minutes for RadeonProRender kernel compilation

Let's run some PBR

root #cd /usr/share/RadeonProRender/tutorials/Bin
root #./22_material_uber64D
RPR Context creation succeeded.
Rendering 22_material_uber_00.png ...
Rendering 22_material_uber_01.png ...
Rendering 22_material_uber_02.png ...
Rendering 22_material_uber_03.png ...
Rendering 22_material_uber_04.png ...
Rendering 22_material_uber_05.png ...
Rendering 22_material_uber_06.png ...
Rendering 22_material_uber_07.png ...
Rendering 22_material_uber_08.png ...
Rendering 22_material_uber_final.png ...

In folder /usr/share/RadeonProRender/tutorials/Bin should be png files with result of rendering. Source code of material creation may be found at /usr/share/RadeonProRender/tutorials/22_material_uber/main.cpp

Image Processing Library

To install Radeon Pro Render Image Filtering library

root #emerge --ask dev-libs/radeon-pro-image-filter

Tutorials

Build

Install the package with the examples USE flag:

root #emerge --ask dev-libs/radeon-pro-image-filter[examples]

Run

All tutorial binaries may be found at /usr/share/RadeonImageFilter/samples/Bin directory. First, let's check ProRender initialize

root #cd /usr/share/RadeonImageFilter
root # mkdir images
root #./samples/Bin/AIDenoiser/AIDenoiser
[MSG  @RIF]: Loaded RIF API version: 1.7.3.0x1c7b817e
INFO: MIOpen version: 2.0.5
[MSG  @RIF]: Loading model: ./models/denoise_c9_ldr.pb
INFO: Memory: required: 0, single node max: 0, requested sum: 0
RIF inference time: 57133 us. Resolution is : 600x800
outputPath=./images/output_0.hdr


In folder /usr/share/RadeonImageFilter/images should be png/hdr files with sample run results


Resources