From Source to ISA: A Trip Down the Shader Compiler Pipeline

Опубликовано 15 июня 2020, 15:05

Shader compilers can be easily confused for dragons — they devour innocent developers, and instead of optimal code, they spew fire. Only the toughest adventurers well versed in the arcane knowledge of how to tame them used to stand a chance.

Until today! Together, we will uncover the dark secrets of the shader compiler and tame this fiery beast. Our journey will take us from the original source code down to the RDNA ISA. We’ll leave no stone unturned when it comes to giving you a deep understanding of compilers. We will arm you with knowledge about control-flow, bindless resource access and scalarization which is directly applicable to problems you’re facing daily at work.

At the frontend, we will look at DXC and how it produces intermediate representations. At the backend, we will discuss the arts of instruction selection, scheduling, and register allocation for the RDNA instruction set.

Throughout the whole trip, we will provide you with practical examples from real-world applications and look at them through the compiler’s eyes. At the end, you will have a solid understanding of your shader compiler and write performant shaders with ease.

***

Visit our website: gpuopen.com
Follow us on Twitter: twitter.com/GPUOpen
Slides: gpuopen.com/wp-content/uploads...

***

Subscribe: bit.ly/Subscribe_to_AMD
Like us on Facebook: bit.ly/AMD_on_Facebook
Follow us on Twitter: bit.ly/AMD_On_Twitter
Follow us on Twitch: Twitch.tv/AMD
Follow us on Linkedin: bit.ly/AMD_on_Linkedin
Follow us on Instagram: bit.ly/AMD_on_Instagram

©2020 Advanced Micro Devices, Inc. AMD, the AMD Arrow Logo, and combinations thereof are trademarks of Advanced Micro Devices, Inc. Other names are for informational purposes only and may be trademarks of their respective owners.