Presentation is loading. Please wait.

Presentation is loading. Please wait.


Similar presentations

Presentation on theme: "GRAPHICS AND COMPUTING GPUS Jehan-François Pâris"— Presentation transcript:


2 Chapter Organization Why bother? Evolution GPU System Architecture Programming GPUs …

3 Why bother? (I) Yesterday's fastest computer was the Sequoia supercomputer –Can crunch 16.32 quadrillion calculations per second (16.32 Petaflops/s). –98,304 compute nodes Each compute nodes is a 16-core PowerPC A2 processor

4 Why bother? (II) Today's fastest computer is the Cray XK7 –Hits 17.59 Petaflops/s on the LINPAC benchmark. –Features 560,640 processors, including 261,632 Nvidia K20x accelerating cores. Supercomputing version of consumer- oriented GK104 CPU

5 Why bother (III) Most techniques developed for high-speed computing end trickling down to mass markets


7 History (I) Up to late 90's –No GPUs –Much simpler VGA controller Consisted of –A memory controller –Display generator + DRAM DRAM was either shared with CPU or private

8 History (I) By 1997 –More complex VGA controllers Incorporated 3D accelerating functions in hardware –Triangle set up and rasterization – Texture mapping and shading

9 Rasterization Converting –An image described in a vector graphics format as a combination of shapes Lines, polygons, letters, … into –A raster image consisting of individual pixels

10 History (II) By 2000 –Single chip graphics processor incorporated nearly all functions of graphics pipeline of high-end workstations Beginning of the end of high-end workstation market –VGA controller was renamed Graphic Processing Units

11 Current trends (I) Graphics processing standards –Well defined APIs – Open GL: Open standard for 3D graphics programming – DirectX: Set of MS multimedia programming interfaces (Direct3D for 3D graphics) Xbox was named after it!

12 Current trends (II) Frequent doubling of GPU speeds –Every 12 to 18 months New paradigm: –Visual computing stands at the intersection graphic processing and parallel computing Can implement novel graphics algorithms Use GPUs for non-conventional applications

13 Two results Triumph of heterogeneous architectures –Combining powers of CPU and GPU GPUs become scalable parallel processors –Moving from hardware-defined pipelining architectures to more flexible programmable architectures

14 From GPGU to CUDA GPGU –General-Purpose computing on GPU –Uses traditional graphics API and graphics pipeline

15 From GPGU to CUDA CUDA –Compute Unified Device Architecture –Parallel computing platform and programming model C/C++ Invented by NVIDIA – Single Program Multiple Data approach


17 Old School Approach CPU North Bridge South Bridge VGA Controller RAM PCI bus Frame buffer UART To VGA display

18 Intel Architecture CPU North Bridge South Bridge DDR2 RAM To display GPU GPU Memory

19 AMD Architecture CPU North Bridge Chipset DDR2 RAM To display GPU GPU Memory

20 Variations Unified Memory Architecture (UMA): –GPU shares RAM with CPU –Lower memory bandwidth, higher latency –Cheap, low-end solution Scalable Link Interconnect: –NVIDIA –Allows multiple GPUs –High-end solution

21 Integrated solutions Integrate CPU and Northbridge Integrate GPU and chipset

22 Game console Similar architectures Architectures evolve over time Objective is to reduce costs while maintaining performance

23 GPU interfaces and drivers GPU attached to CPU via PCI-Express –Replaces older AGP Interfaces such as OpenGL and Direct3D use the GPU as a coprocessor –Send commands, programs and data to GPU through a specific GPU device driver They are often buggy!

24 Graphics logical pipeline Vertex Shader Geometry Shader Setup & Raster Pixel Shader Raster & Merger These functions must be mapped into a programmable GPU Input Ass'er

25 Basic Unified GPU Architecture Programmable processor array –Tightly integrated with fixed-function processors for texture filtering, rasterization, raster operations –Emphasis in on very high level of parallelism

26 Example architecture Tesla architecture (NVIDIA Geoforce 8800) 116 streaming processors (SP) cores –Organized as 14 multithreaded streaming multiprocessors (SM) Each SP core –Manages 96 concurrent threads Thread state are maintained by hardware –Connects with four 64-bit DRAM partitions

27 Example architecture Each SM has –8 SP cores –2 special function units –Separate caches for instructions and constants –A multithreaded instruction unit –Shared memory (NUMA?)

28 PROGRAMMING GPUS Will focus on parallel computing applications

29 Key idea Must decompose problem into set of parallel computations –Ideally two-level to match GPU organization

30 Example Data are in big array Small array Tiny

31 CUDA CUDA programs are written in C Provides three abstractions –Hierarchy of thread groups –Shared memory –Barrier synchronization

32 Barrier synchronization Barriers let threads –Wait for completion of a computation step by other cores so they can Exchange results Start next step

33 Example Tiny Barrier = Wait for each other Exchange partial results Tiny Barrier = Wait for each other Exchange partial results Tiny

34 Big fallacies GPUs –Not good for general computation –Cannot run double precision arithmetic –Do not do floating point correctly Cannot speedup O(n) algorithms

Download ppt "GRAPHICS AND COMPUTING GPUS Jehan-François Pâris"

Similar presentations

Ads by Google