CSC/ECE 506 Spring 2014/2b hd - Revision history

Dsbhat: /* Introduction to GPU */

2014-02-11T04:19:45Z

Introduction to GPU

← Older revision		Revision as of 04:19, 11 February 2014
Line 5:		Line 5:

	==Introduction to GPU==		==Introduction to GPU==
	Graphical Processing Unit (GPU) is an hardware accelerator used for graphical computation necessary to display the image on the screen. The GPUs were designed initially to alleviate the CPU work load. GPUs are characterized by high programmability, fast floating point parallel computation, minimal dependency on a CPU, large data sets and cost efficiency. The immense capability of the GPU to perform high speed computations involving vectors and matrices has made them suitable for non-graphical applications as well. <ref name= GPU vs CPU by NVIDIA> GPU vs CPU by NVIDIA. URL http://cs.utsa.edu/~qitian/seminar/Spring11/03_04_11/GPU.pdf </ref>		Graphical Processing Unit (GPU) is an hardware accelerator used for graphical computation, necessary to display the image on the screen. The GPUs were designed initially to alleviate the CPU work load. GPUs are characterized by high programmability, fast floating point parallel computation, minimal dependency on a CPU, large data sets and cost efficiency. The immense capability of the GPU to perform high speed computations involving vectors and matrices has made them suitable for non-graphical applications as well. <ref name= GPU vs CPU by NVIDIA> GPU vs CPU by NVIDIA. URL http://cs.utsa.edu/~qitian/seminar/Spring11/03_04_11/GPU.pdf </ref>

	The various functionalities carried out by a GPU include computations related to 3D computer graphics, texture mapping, geometric and vertex calculations and Rasterization. This makes it a best choice processor for applications such as Video decoding, Molecular dynamics, Quantum Chemistry, Material Science, Visualization and Docking Software, Bioinformatics, Numerical Analytics, Physics, Defense and Intelligence, Computational Finance, Computational Fluid Dynamics, Computational Structural mechanics, Computer Aided Design, Animation and Modeling, Color correction and grain management, Oil and Gas.		The various functionalities carried out by a GPU include computations related to 3D computer graphics, texture mapping, geometric and vertex calculations and Rasterization. This makes it a best choice processor for applications such as Video decoding, Molecular dynamics, Quantum Chemistry, Material Science, Visualization and Docking Software, Bioinformatics, Numerical Analytics, Physics, Defense and Intelligence, Computational Finance, Computational Fluid Dynamics, Computational Structural mechanics, Computer Aided Design, Animation and Modeling, Color correction and grain management, Oil and Gas.

Dsbhat: /* Introduction to GPU */

2014-02-11T04:18:24Z

Introduction to GPU

← Older revision		Revision as of 04:18, 11 February 2014
Line 5:		Line 5:

	==Introduction to GPU==		==Introduction to GPU==
	Graphical Processing Unit (GPU) is an hardware accelerator used for graphical computation necessary to display the image on the screen. The GPUs were designed initially to alleviate the CPU work load. GPUs are characterized by high programmability, fast floating point parallel computation, minimal dependency on a CPU, large data sets and cost efficiency. The immense capability of the GPU to perform high speed computations involving vectors and matrices has made them suitable for non-graphical applications as well.		Graphical Processing Unit (GPU) is an hardware accelerator used for graphical computation necessary to display the image on the screen. The GPUs were designed initially to alleviate the CPU work load. GPUs are characterized by high programmability, fast floating point parallel computation, minimal dependency on a CPU, large data sets and cost efficiency. The immense capability of the GPU to perform high speed computations involving vectors and matrices has made them suitable for non-graphical applications as well. <ref name= GPU vs CPU by NVIDIA> GPU vs CPU by NVIDIA. URL http://cs.utsa.edu/~qitian/seminar/Spring11/03_04_11/GPU.pdf </ref>

	The various functionalities carried out by a GPU include computations related to 3D computer graphics, texture mapping, geometric and vertex calculations and Rasterization. This makes it a best choice processor for applications such as Video decoding, Molecular dynamics, Quantum Chemistry, Material Science, Visualization and Docking Software, Bioinformatics, Numerical Analytics, Physics, Defense and Intelligence, Computational Finance, Computational Fluid Dynamics, Computational Structural mechanics, Computer Aided Design, Animation and Modeling, Color correction and grain management, Oil and Gas.		The various functionalities carried out by a GPU include computations related to 3D computer graphics, texture mapping, geometric and vertex calculations and Rasterization. This makes it a best choice processor for applications such as Video decoding, Molecular dynamics, Quantum Chemistry, Material Science, Visualization and Docking Software, Bioinformatics, Numerical Analytics, Physics, Defense and Intelligence, Computational Finance, Computational Fluid Dynamics, Computational Structural mechanics, Computer Aided Design, Animation and Modeling, Color correction and grain management, Oil and Gas.

Dsbhat: /* GPU VS CPU */

2014-02-11T04:05:54Z

GPU VS CPU

← Older revision		Revision as of 04:05, 11 February 2014
Line 13:		Line 13:

	===GPU VS CPU===		===GPU VS CPU===
	A GPU is made up of many tiny processing cores which are very efficient, easily programmable and highly suitable for parallel computation. In contrast, a CPU has few cores optimized for sequential serial processing. The diagram below shows how the processing cores are arranged in CPU and GPU <ref name= GPU vs CPU by NVIDIA> GPU vs CPU by NVIDIA. URL http://www.nvidia.com/object/what-is-gpu-computing.html </ref>		A GPU is made up of many tiny processing cores which are very efficient, easily programmable and highly suitable for parallel computation. In contrast, a CPU has few cores optimized for sequential serial processing. The cache size and flow control units are however of smaller size when compared to the CPU. This indicates that during the fabrication, bulk of the transistors are devoted for data processing. The diagram below shows how the processing cores are arranged in CPU and GPU <ref name= GPU vs CPU by NVIDIA> GPU vs CPU by NVIDIA. URL http://www.nvidia.com/object/what-is-gpu-computing.html </ref>

	[[File:GPUvsCPU.jpg\|center\|Figure1. GPU vs CPU ]]		[[File:GPUvsCPU.jpg\|center\|Figure1. GPU vs CPU ]]

Dsbhat: /* GPU VS CPU */

2014-02-11T04:00:25Z

GPU VS CPU

← Older revision		Revision as of 04:00, 11 February 2014
Line 17:		Line 17:
	[[File:GPUvsCPU.jpg\|center\|Figure1. GPU vs CPU ]]		[[File:GPUvsCPU.jpg\|center\|Figure1. GPU vs CPU ]]

	<center>Figure 1. GPU vs CPU <ref name= GPUCPU> GPU vs CPU. URL http://~~ixbtlabs~~.com/~~articles3/video~~/cuda-1-~~p1.html~~ </ref> </center>		<center>Figure 1. GPU vs CPU <ref name= GPUCPU> GPU vs CPU. URL http://docs.nvidia.com/cuda/cuda-c-programming-guide/ </ref> </center>

	==GPU Architecture==		==GPU Architecture==

Dsbhat: /* GPU VS CPU */

2014-02-11T03:56:12Z

GPU VS CPU

← Older revision		Revision as of 03:56, 11 February 2014
Line 17:		Line 17:
	[[File:GPUvsCPU.jpg\|center\|Figure1. GPU vs CPU ]]		[[File:GPUvsCPU.jpg\|center\|Figure1. GPU vs CPU ]]

	<center>Figure 1. GPU vs CPU~~</center>~~ <ref name= GPUCPU> GPU vs CPU. URL http://ixbtlabs.com/articles3/video/cuda-1-p1.html </ref>		<center>Figure 1. GPU vs CPU <ref name= GPUCPU> GPU vs CPU. URL http://ixbtlabs.com/articles3/video/cuda-1-p1.html </ref> </center>

	==GPU Architecture==		==GPU Architecture==

Dsbhat: /* GPU VS CPU */

2014-02-11T03:55:50Z

GPU VS CPU

← Older revision		Revision as of 03:55, 11 February 2014
Line 15:		Line 15:
	A GPU is made up of many tiny processing cores which are very efficient, easily programmable and highly suitable for parallel computation. In contrast, a CPU has few cores optimized for sequential serial processing. The diagram below shows how the processing cores are arranged in CPU and GPU <ref name= GPU vs CPU by NVIDIA> GPU vs CPU by NVIDIA. URL http://www.nvidia.com/object/what-is-gpu-computing.html </ref>		A GPU is made up of many tiny processing cores which are very efficient, easily programmable and highly suitable for parallel computation. In contrast, a CPU has few cores optimized for sequential serial processing. The diagram below shows how the processing cores are arranged in CPU and GPU <ref name= GPU vs CPU by NVIDIA> GPU vs CPU by NVIDIA. URL http://www.nvidia.com/object/what-is-gpu-computing.html </ref>

	[[File:GPUvsCPU.jpg\|center\|Figure1. GPU vs CPU ]]~~</center>~~		[[File:GPUvsCPU.jpg\|center\|Figure1. GPU vs CPU ]]

	<center>Figure 1. GPU vs CPU</center> <ref name= GPUCPU> GPU vs CPU. URL http://ixbtlabs.com/articles3/video/cuda-1-p1.html </ref>		<center>Figure 1. GPU vs CPU</center> <ref name= GPUCPU> GPU vs CPU. URL http://ixbtlabs.com/articles3/video/cuda-1-p1.html </ref>

Dsbhat: /* GPU VS CPU */

2014-02-11T03:55:15Z

GPU VS CPU

← Older revision		Revision as of 03:55, 11 February 2014
Line 15:		Line 15:
	A GPU is made up of many tiny processing cores which are very efficient, easily programmable and highly suitable for parallel computation. In contrast, a CPU has few cores optimized for sequential serial processing. The diagram below shows how the processing cores are arranged in CPU and GPU <ref name= GPU vs CPU by NVIDIA> GPU vs CPU by NVIDIA. URL http://www.nvidia.com/object/what-is-gpu-computing.html </ref>		A GPU is made up of many tiny processing cores which are very efficient, easily programmable and highly suitable for parallel computation. In contrast, a CPU has few cores optimized for sequential serial processing. The diagram below shows how the processing cores are arranged in CPU and GPU <ref name= GPU vs CPU by NVIDIA> GPU vs CPU by NVIDIA. URL http://www.nvidia.com/object/what-is-gpu-computing.html </ref>

	[[File:GPUvsCPU.jpg\|center\|Figure1. GPU vs CPU ]]~~<ref name= GPUCPU> GPU vs CPU. URL http://ixbtlabs.com/articles3/video/cuda-1-p1.html </ref>~~ </center>		[[File:GPUvsCPU.jpg\|center\|Figure1. GPU vs CPU ]]</center>

	<center>Figure 1. GPU vs CPU</center>		<center>Figure 1. GPU vs CPU</center> <ref name= GPUCPU> GPU vs CPU. URL http://ixbtlabs.com/articles3/video/cuda-1-p1.html </ref>

	==GPU Architecture==		==GPU Architecture==

Dsbhat: /* GPU VS CPU */

2014-02-11T03:54:21Z

GPU VS CPU

← Older revision		Revision as of 03:54, 11 February 2014
Line 15:		Line 15:
	A GPU is made up of many tiny processing cores which are very efficient, easily programmable and highly suitable for parallel computation. In contrast, a CPU has few cores optimized for sequential serial processing. The diagram below shows how the processing cores are arranged in CPU and GPU <ref name= GPU vs CPU by NVIDIA> GPU vs CPU by NVIDIA. URL http://www.nvidia.com/object/what-is-gpu-computing.html </ref>		A GPU is made up of many tiny processing cores which are very efficient, easily programmable and highly suitable for parallel computation. In contrast, a CPU has few cores optimized for sequential serial processing. The diagram below shows how the processing cores are arranged in CPU and GPU <ref name= GPU vs CPU by NVIDIA> GPU vs CPU by NVIDIA. URL http://www.nvidia.com/object/what-is-gpu-computing.html </ref>

	[[File:GPUvsCPU.jpg\|center\|Figure1. GPU vs CPU ]]		[[File:GPUvsCPU.jpg\|center\|Figure1. GPU vs CPU ]]<ref name= GPUCPU> GPU vs CPU. URL http://ixbtlabs.com/articles3/video/cuda-1-p1.html </ref> </center>

	<center>Figure 1. GPU vs CPU</center>		<center>Figure 1. GPU vs CPU</center>

Dsbhat at 03:42, 11 February 2014

2014-02-11T03:42:55Z

← Older revision		Revision as of 03:42, 11 February 2014
Line 5:		Line 5:

	==Introduction to GPU==		==Introduction to GPU==
	Graphical Processing Unit (GPU) is a hardware accelerator characterized by high programmability, fast floating point parallel computation, minimal dependency on a CPU, large data sets and cost efficiency~~. The GPUs were designed initially to perform intensive computations related to graphics in order to alleviate the CPU work load~~. The immense capability of the GPU to perform high speed computations involving vectors and matrices has made them suitable for non-graphical applications as well.		Graphical Processing Unit (GPU) is an hardware accelerator used for graphical computation necessary to display the image on the screen. The GPUs were designed initially to alleviate the CPU work load. GPUs are characterized by high programmability, fast floating point parallel computation, minimal dependency on a CPU, large data sets and cost efficiency. The immense capability of the GPU to perform high speed computations involving vectors and matrices has made them suitable for non-graphical applications as well.

	The various functionalities carried out by a GPU include computations related to 3D computer graphics, texture mapping, geometric and vertex calculations and Rasterization. This makes it a best choice processor for applications such as Video decoding, Molecular dynamics, Quantum Chemistry, Material Science, Visualization and Docking Software, Bioinformatics, Numerical Analytics, Physics, Defense and Intelligence, Computational Finance, Computational Fluid Dynamics, Computational Structural mechanics, Computer Aided Design, Animation and Modeling, Color correction and grain management, Oil and Gas.		The various functionalities carried out by a GPU include computations related to 3D computer graphics, texture mapping, geometric and vertex calculations and Rasterization. This makes it a best choice processor for applications such as Video decoding, Molecular dynamics, Quantum Chemistry, Material Science, Visualization and Docking Software, Bioinformatics, Numerical Analytics, Physics, Defense and Intelligence, Computational Finance, Computational Fluid Dynamics, Computational Structural mechanics, Computer Aided Design, Animation and Modeling, Color correction and grain management, Oil and Gas.

Dsbhat: /* Questions */

2014-02-03T23:04:32Z

Questions

← Older revision		Revision as of 23:04, 3 February 2014
Line 259:		Line 259:

	==Questions==		==Questions==
	*Q1. ~~Suppose~~ a ~~GPU from AMD supporting the Cayman architecture has 24 SIMD cores and 1728 processing elements in total. How many processing elements are present in 1 core~~?		*Q1. What makes GPU a best choice for graphical operations?

	*Q2. ~~Name~~ the ~~programming model developed by NVidia~~		*Q2. What is the difference between a CPU and a GPU?

	*Q3. ~~What is~~ the ~~syntax used to create N kernels~~ in ~~CUDA C~~		*Q3. Suppose a GPU from AMD supporting the Cayman architecture has 24 SIMD cores and 1728 processing elements in total. How many processing elements are present in 1 core?

	*Q4. ~~In CUDA, what directive is used to denote that~~ the ~~code~~ is ~~a parallel kernel?~~		*Q4. Explain how the GPU hardware is supportive in achieving data parallelism.

	*Q5. ~~Who was responsible for~~ the ~~initial development of OpenGL?~~		*Q5. Name the programming model developed by NVidia

	*Q6. What is the ~~primary difference between OpenCL and OpenGL application operations~~?		*Q6. What is the importance of CUDA language?

	*Q7. What ~~other types of parallelism are present~~ in ~~GPU architecture?~~		*Q7. What is the syntax used to create N kernels in CUDA C

	*Q8. ~~What advantages do GPUs have over CPUs~~?		*Q8. In CUDA, what directive is used to denote that the code is a parallel kernel?

	*Q9. In OpenGL, which tasks are assigned to CPUs and which tasks are handled by GPUs?		*Q9. How is parallel programming achieved using CUDA language?

			*Q10. Who was responsible for the initial development of OpenGL?

			*Q11. What is the primary difference between OpenCL and OpenGL application operations?

			*Q12. What other types of parallelism are present in GPU architecture?

			*Q13. What advantages do GPUs have over CPUs?

			*Q14. In OpenGL, which tasks are assigned to CPUs and which tasks are handled by GPUs?

			*Q15. What are the main functions of kernels in OpenCL?

	*Q10. What are the main functions of kernels in OpenCL?
	==References==		==References==
	<references/>		<references/>