CUDA (Compute Unified Device Dr. Bharathwaj Bharath Muthuswamy - PowerPoint PPT Presentation

CUDA (Compute Unified Device Architecture) and OpenCL (Open Compute Language): Programming GPUs CUDA (Compute Unified Device Dr. Bharathwaj “Bharath” Muthuswamy Architecture) and OpenCL (Open Compute About me... Language): Programming GPUs Goals and Organization Conceptual Overview of the Project Understanding the Graphics Pipeline Dr. Bharathwaj “Bharath” Muthuswamy Part I: CUDA Part II: OpenCL Visiting Assistant Professor in Computer Science DTCNN and OpenCL The College of New Jersey Conclusion and Q/A bharath.berkeley@gmail.com February 20th 2018 1

CUDA (Compute About me... Unified Device Architecture) and OpenCL (Open Compute Language): Programming GPUs ◮ BS (2002), MS (2005), PhD (2009) in EECS from the Dr. Bharathwaj “Bharath” University of California, Berkeley Muthuswamy (advisors: Dr. Leon O. Chua, Dr. Pravin P. Varaiya) About me... Goals and Organization Conceptual Overview of the Project Understanding the Graphics Pipeline Part I: CUDA Part II: OpenCL DTCNN and OpenCL Conclusion and Q/A 2

CUDA (Compute About me... Unified Device Architecture) and OpenCL (Open Compute Language): Programming GPUs ◮ BS (2002), MS (2005), PhD (2009) in EECS from the Dr. Bharathwaj “Bharath” University of California, Berkeley Muthuswamy (advisors: Dr. Leon O. Chua, Dr. Pravin P. Varaiya) About me... ◮ For my MS, I worked on biomimetic bipedal robotics Goals and Organization using Central Pattern Generators (I did not work on this Conceptual Overview of the Project after 2006) Understanding the ◮ For my PhD, my primary contribution was designing, Graphics Pipeline implementing and rigorously proving the existence of Part I: CUDA Part II: OpenCL chaos in the Muthuswamy-Chua system (circuit): an DTCNN and OpenCL inductor-capacitor-memristor circuit in series (parallel) Conclusion and Q/A 2

CUDA (Compute About me... Unified Device Architecture) and OpenCL (Open Compute Language): Programming GPUs ◮ BS (2002), MS (2005), PhD (2009) in EECS from the Dr. Bharathwaj “Bharath” University of California, Berkeley Muthuswamy (advisors: Dr. Leon O. Chua, Dr. Pravin P. Varaiya) About me... ◮ For my MS, I worked on biomimetic bipedal robotics Goals and Organization using Central Pattern Generators (I did not work on this Conceptual Overview of the Project after 2006) Understanding the ◮ For my PhD, my primary contribution was designing, Graphics Pipeline implementing and rigorously proving the existence of Part I: CUDA Part II: OpenCL chaos in the Muthuswamy-Chua system (circuit): an DTCNN and OpenCL inductor-capacitor-memristor circuit in series (parallel) Conclusion and Q/A ◮ Areas of interest: ◮ Computer vision and Quantum Computing ◮ Nonlinear Dynamics (Circuits). Specifically: chaotic circuits and memristors ◮ Embedded (FPGA) Systems and Education 2

CUDA (Compute Project goal, Presentation goal and Organization Unified Device Architecture) and OpenCL (Open Compute Language): Programming GPUs ◮ Goal of my research: Implement a model for the Dr. Bharathwaj mammalian visual pathway on a heterogeneous platform “Bharath” Muthuswamy using Discrete Time Cellular Nonlinear Networks (DTCNN). About me... Goals and Organization Conceptual Overview of the Project Understanding the Graphics Pipeline Part I: CUDA Part II: OpenCL DTCNN and OpenCL Conclusion and Q/A 3

CUDA (Compute Project goal, Presentation goal and Organization Unified Device Architecture) and OpenCL (Open Compute Language): Programming GPUs ◮ Goal of my research: Implement a model for the Dr. Bharathwaj mammalian visual pathway on a heterogeneous platform “Bharath” Muthuswamy using Discrete Time Cellular Nonlinear Networks (DTCNN). About me... Goals and Organization ◮ Application: Robust recognition of hand-drawn Conceptual Overview electronic circuit diagrams of the Project Understanding the ◮ Goal of this talk: Discuss GPU programming strategies Graphics Pipeline Part I: CUDA (CUDA and OpenCL) Part II: OpenCL DTCNN and OpenCL Conclusion and Q/A 3

CUDA (Compute Project goal, Presentation goal and Organization Unified Device Architecture) and OpenCL (Open Compute Language): Programming GPUs ◮ Goal of my research: Implement a model for the Dr. Bharathwaj mammalian visual pathway on a heterogeneous platform “Bharath” Muthuswamy using Discrete Time Cellular Nonlinear Networks (DTCNN). About me... Goals and Organization ◮ Application: Robust recognition of hand-drawn Conceptual Overview electronic circuit diagrams of the Project Understanding the ◮ Goal of this talk: Discuss GPU programming strategies Graphics Pipeline Part I: CUDA (CUDA and OpenCL) Part II: OpenCL ◮ Organization: DTCNN and OpenCL Conclusion and Q/A 3

CUDA (Compute Project goal, Presentation goal and Organization Unified Device Architecture) and OpenCL (Open Compute Language): Programming GPUs ◮ Goal of my research: Implement a model for the Dr. Bharathwaj mammalian visual pathway on a heterogeneous platform “Bharath” Muthuswamy using Discrete Time Cellular Nonlinear Networks (DTCNN). About me... Goals and Organization ◮ Application: Robust recognition of hand-drawn Conceptual Overview electronic circuit diagrams of the Project Understanding the ◮ Goal of this talk: Discuss GPU programming strategies Graphics Pipeline Part I: CUDA (CUDA and OpenCL) Part II: OpenCL ◮ Organization: DTCNN and OpenCL ◮ Conceptual Overview of the Project Conclusion and Q/A 3

CUDA (Compute Project goal, Presentation goal and Organization Unified Device Architecture) and OpenCL (Open Compute Language): Programming GPUs ◮ Goal of my research: Implement a model for the Dr. Bharathwaj mammalian visual pathway on a heterogeneous platform “Bharath” Muthuswamy using Discrete Time Cellular Nonlinear Networks (DTCNN). About me... Goals and Organization ◮ Application: Robust recognition of hand-drawn Conceptual Overview electronic circuit diagrams of the Project Understanding the ◮ Goal of this talk: Discuss GPU programming strategies Graphics Pipeline Part I: CUDA (CUDA and OpenCL) Part II: OpenCL ◮ Organization: DTCNN and OpenCL ◮ Conceptual Overview of the Project Conclusion and Q/A ◮ Understanding the graphics pipeline 3

CUDA (Compute Project goal, Presentation goal and Organization Unified Device Architecture) and OpenCL (Open Compute Language): Programming GPUs ◮ Goal of my research: Implement a model for the Dr. Bharathwaj mammalian visual pathway on a heterogeneous platform “Bharath” Muthuswamy using Discrete Time Cellular Nonlinear Networks (DTCNN). About me... Goals and Organization ◮ Application: Robust recognition of hand-drawn Conceptual Overview electronic circuit diagrams of the Project Understanding the ◮ Goal of this talk: Discuss GPU programming strategies Graphics Pipeline Part I: CUDA (CUDA and OpenCL) Part II: OpenCL ◮ Organization: DTCNN and OpenCL ◮ Conceptual Overview of the Project Conclusion and Q/A ◮ Understanding the graphics pipeline ◮ Part I: CUDA - Hello, world; SAXPY 3

CUDA (Compute Project goal, Presentation goal and Organization Unified Device Architecture) and OpenCL (Open Compute Language): Programming GPUs ◮ Goal of my research: Implement a model for the Dr. Bharathwaj mammalian visual pathway on a heterogeneous platform “Bharath” Muthuswamy using Discrete Time Cellular Nonlinear Networks (DTCNN). About me... Goals and Organization ◮ Application: Robust recognition of hand-drawn Conceptual Overview electronic circuit diagrams of the Project Understanding the ◮ Goal of this talk: Discuss GPU programming strategies Graphics Pipeline Part I: CUDA (CUDA and OpenCL) Part II: OpenCL ◮ Organization: DTCNN and OpenCL ◮ Conceptual Overview of the Project Conclusion and Q/A ◮ Understanding the graphics pipeline ◮ Part I: CUDA - Hello, world; SAXPY ◮ Part II: OpenCL - Hello, world 3

CUDA (Compute Project goal, Presentation goal and Organization Unified Device Architecture) and OpenCL (Open Compute Language): Programming GPUs ◮ Goal of my research: Implement a model for the Dr. Bharathwaj mammalian visual pathway on a heterogeneous platform “Bharath” Muthuswamy using Discrete Time Cellular Nonlinear Networks (DTCNN). About me... Goals and Organization ◮ Application: Robust recognition of hand-drawn Conceptual Overview electronic circuit diagrams of the Project Understanding the ◮ Goal of this talk: Discuss GPU programming strategies Graphics Pipeline Part I: CUDA (CUDA and OpenCL) Part II: OpenCL ◮ Organization: DTCNN and OpenCL ◮ Conceptual Overview of the Project Conclusion and Q/A ◮ Understanding the graphics pipeline ◮ Part I: CUDA - Hello, world; SAXPY ◮ Part II: OpenCL - Hello, world ◮ Current work: OpenCL specification of DTCNN 3