US 11,868,759 B2
Shader source code performance prediction
Amit Ben-Moshe, Boston, MA (US); and Ian Charles Colbert, San Diego, CA (US)
Assigned to Advanced Micro Devices, Inc., Santa Clara, CA (US)
Filed by Advanced Micro Devices, Inc., Santa Clara, CA (US)
Filed on Dec. 8, 2021, as Appl. No. 17/545,801.
Prior Publication US 2023/0176847 A1, Jun. 8, 2023
Int. Cl. G06F 8/65 (2018.01); G06T 15/00 (2011.01); G06N 20/00 (2019.01); G06F 8/51 (2018.01); G06F 8/41 (2018.01)
CPC G06F 8/65 (2013.01) [G06F 8/443 (2013.01); G06F 8/51 (2013.01); G06N 20/00 (2019.01); G06T 15/005 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method comprising:
receiving, using hardware of one or more computing devices, an update to shader source code for implementing a shader;
after the update to the shader source code and using the hardware, compiling the shader source code with the update to generate a representation of the shader;
providing the representation of the shader as input to one or more machine learning models trained based on historical data describing performance of one or more other shaders on a processing unit;
receiving, as an output from the one or more machine learning models, a prediction of performance of the shader on the processing unit based on the update to the shader source code; and
outputting, using the hardware, feedback about the update, the feedback including the prediction of performance of the shader.