GRAPHICS PROCESSING UNIT(GPU)

A PRESENTATION BY

P. JAFAR KHAN160040686

INTRODUCTION

• THE GRAPHICS PROCESSING UNIT (GPU) HAS BECOME AN INTEGRAL PART OF TODAY’S MAINSTREAM COMPUTING SYSTEMS. OVER THE PAST SIX YEARS, THERE HAS BEEN A MARKED INCREASE IN THE PERFORMANCE AND CAPABILITIES OF GPUS.• GPU IS A GRAPHICAL PROCESSING UNIT WHICH ENABLES YOU RUN

HIGH DEFINITIONS GRAPHICS ON YOUR PC, WHICH ARE THE EMEND OF MODERN COMPUTING. LIKE THE CPU (CENTRAL PROCESSING UNIT), IT IS A SINGLE-CHIP PROCESSOR. THE GPU HAS HUNDREDS OF CORES AS COMPARED TO THE 4 OR 8 IN THE LATEST CPUS. THE PRIMARY JOB OF THE GPU IS TO COMPUTE 3D FUNCTIONS.

PICTURES OF THE GPUNvidia Geforce GTX 1070 FTW 8 GB

AMD Radeon R9 295X2

GPU ARCHITECTURE• CONTROL HARDWARE DOMINATES PROCESSORS• COMPLEX, DIFFICULT TO BUILD AND VERIFY• TAKES SUBSTANTIAL FRACTION OF DIE SCALES POORLY• PAY FOR MAX THROUGHPUT, SUSTAIN AVERAGE THROUGHPUT• QUADRATIC DEPENDENCY CHECKING• CONTROL HARDWARE DOESN’T DO ANY MATH!• OVER THE PAST FEW YEARS, THE GPU HAS EVOLVED FROM A

FIXED-FUNCTION SPECIAL-PURPOSE PROCESSOR INTO A FULL-FLEDGED PARALLEL PROGRAMMABLE PROCESSOR WITH ADDITIONAL FIXED-FUNCTION SPECIAL-PURPOSE FUNCTIONALITY.

GPU ARCHITECTURE• THE GRAPHICS PIPELINE

- THE INPUT TO THE GPU IS A LIST OF GEOMETRIC PRIMITIVES, TYPICALLY TRIANGLES, IN A 3-D WORLD COORDINATE SYSTEM. THROUGH MANY STEPS, VERTEX OPERATIONS: THE INPUT PRIMITIVES ARE FORMED FROM INDIVIDUAL VERTICES. EACH VERTEX MUST BE TRANSFORMED INTO SCREEN SPACE AND SHADED, TYPICALLY THROUGH COMPUTING THEIR INTERACTION WITH THE LIGHTS IN THE SCENE. BECAUSE TYPICAL SCENES HAVE TENS TO HUNDREDS OF THOUSANDS OF VERTICES, AND EACH VERTEX CAN BE COMPUTED INDEPENDENTLY, THIS STAGE IS WELL SUITED FOR PARALLEL HARDWARE.

GPU ARCHITECTURE

GPU ARCHITECTURE• EVOLUTION OF GPU ARCHITECTURE

- THE FIXED-FUNCTION PIPELINE LACKED THE GENERALITY TO EFFICIENTLY EXPRESS MORE COMPLICATED SHADING AND LIGHTING OPERATIONS THAT ARE ESSENTIAL FOR COMPLEX EFFECTS. THE KEY STEP WAS REPLACING THE FIXED-FUNCTION PER-VERTEX AND PER-FRAGMENT OPERATIONS WITH USER-SPECIFIED PROGRAMS RUN ON EACH VERTEX AND FRAGMENT. OVER THE PAST SIX YEARS, THESE VERTEX PROGRAMS AND FRAGMENT PROGRAMS HAVE BECOME INCREASINGLY MORE CAPABLE, WITH LARGER LIMITS ON THEIR SIZE AND RESOURCE CONSUMPTION, WITH MORE FULLY FEATURED INSTRUCTION SETS, AND WITH MORE FLEXIBLE CONTROL-FLOW OPERATIONS.

GPU ARCHITECTURE• ARCHITECTURE OF MODERN GPU

- WE NOTED THAT THE GPU IS BUILT FOR DIFFERENT APPLICATION DEMANDS THAN THE CPU: LARGE PARALLEL COMPUTATION REQUIREMENTS WITH AN EMPHASIS ON THROUGHPUT RATHER THAN LATENCY. CONSEQUENTLY, THE ARCHITECTURE OF THE GPU HAS PROGRESSED IN A DIFFERENT DIRECTION THAN THAT OF THE CPU.

- THE CPU DIVIDES THE PIPELINE IN TIME, APPLYING ALL RESOURCES IN THE PROCESSOR TO EACH STAGE IN TURN. GPUS HAVE HISTORICALLY TAKEN A DIFFERENT APPROACH. THE GPU DIVIDES THE RESOURCES OF THE PROCESSOR AMONG THE DIFFERENT STAGES, SUCH THAT THE PIPELINE IS DIVIDED IN SPACE, NOT TIME. THE PART OF THE PROCESSOR WORKING ON ONE STAGE FEEDS ITS OUTPUT DIRECTLY INTO A DIFFERENT PART THAT WORKS ON THE NEXT STAGE.

GPU ARCHITECTURE

LATEST TECHNOLOGIES USED

• NVIDIA– TESLA HPC SPECIFIC GPUS HAVE EVOLVED FROM

GEFORCE SERIES• AMD

– FIRE STREAM HPC SPECIFIC GPUS HAVE EVOLVED FROM (ATI) RADEON SERIES• INTEL

– KNIGHTS CORNER MANY-CORE X86 CHIP IS LIKE HYBRID BETWEEN A GPU AND MANY-CORE CPU

CHARACTERISTICS OF GRAPHICS IN GPU

• LARGE COMPUTATIONAL REQUIREMENTS• MASSIVE PARALLELISM• GRAPHICS PIPELINE DESIGNED FOR INDEPENDENT OPERATIONS• LONG LATENCIES TOLERABLE• DEEP, FEED-FORWARD PIPELINES• HACKS ARE OK—CAN TOLERATE LACK OF ACCURACY• GPUS ARE GOOD AT PARALLEL, ARITHMETICALLY INTENSE,

STREAMING-MEMORY PROBLEMS

ASSUMPTIONS AND LIMITATIONS IN GPU

• IN ADDITION TO QUERY PROCESSING, LARGE WEB SEARCH ENGINES NEED TO PERFORM MANY OTHER OPERATIONS INCLUDING WEB CRAWL-ING, INDEX BUILDING, AND DATA MINING STEPS FOR TASKS SUCH AS LINK ANALYSIS AND SPAM AND DUPLICATE DETECTION. WE FOCUS HERE ON QUERY PROCESSING AND IN PARTICULAR ON ONE PHASE OF THIS STEP AS EXPLAINED FURTHER BELOW. WE BELIEVE THAT THIS PART IS SUITABLE FOR IMPLEMENTATION ON GPUS AS IT IS FAIRLY SIMPLE IN STRUCTURE BUT NONETHELESS CONSUMES A DISPROPORTIONATE AMOUNT OF THE OVERALL SYSTEM RESOURCES. IN CONTRAST, WE DO NOT THINK THAT IMPLEMENTATION OF A COMPLETE SEARCH ENGINE ON A GPU IS CURRENTLY REALISTIC.

GPU COMPUTING

• THE GPU PROGRAMMING MODEL- THE PROGRAMMABLE UNITS OF THE GPU FOLLOW A SINGLE

PROGRAM MULTIPLE-DATA (SPMD) PROGRAMMING MODEL. FOR EFFICIENCY, THE GPU PROCESSES MANY ELEMENTS (VERTICES OR FRAGMENTS) IN PARALLEL USING THE SAME PROGRAM. EACH ELEMENT IS INDEPENDENT FROM THE OTHER ELEMENTS, AND IN THE BASE PROGRAMMING MODEL, ELEMENTS CANNOT COMMUNICATE WITH EACH OTHER. ALL GPU PROGRAMS MUST BE STRUCTURED IN THIS WAY: MANY PARALLEL ELEMENTS, EACH PROCESSED IN PARALLEL BY A SINGLE PROGRAM.

GPU COMPUTING• GENERAL-PURPOSE COMPUTING ON THE GPU

- STEPS TO SHOW THE SIMPLER AND DIRECT WAY THAT TODAY’S GPU COMPUTING APPLICATIONS ARE WRITTEN.1. PROGRAMMING A GPU FOR GRAPHICS: WE BEGIN WITH THE SAME GPU PIPELINE THAT WE

DESCRIBED IN SECTION II, CONCENTRATING ON THE PROGRAMMABLE ASPECTS OF THIS PIPELINE.

2. THE PROGRAMMER SPECIFIES GEOMETRY THAT COVERS A REGION ON THE SCREEN. THE RASTERIZER GENERATES A FRAGMENT AT EACH PIXEL LOCATION COVERED BY THAT GEOMETRY.

3. EACH FRAGMENT IS SHADED BY THE FRAGMENT PROGRAM.4. THE FRAGMENT PROGRAM COMPUTES THE VALUE OF THE FRAGMENT BY A COMBINATION OF

MATH OPERATIONS AND GLOBAL MEMORY READS FROM A GLOBAL BTEXTURE [MEMORY].5. THE RESULTING IMAGE CAN THEN BE USED AS TEXTURE ON FUTURE PASSES THROUGH THE

GRAPHICS PIPELINE.

GPU VIRTUALIZATION TAXONOMY

• WE OBSERVE THAT DIFFERENT USE CASES WEIGHT THE CRITERIA DIFFERENTLY—FOR EXAMPLE A VDI DEPLOYMENT VALUES HIGH VM-TO-GPU CONSOLIDATION RATIOS (E.G., MULTIPLEXING) WHILE A CONSUMER RUNNING A VM TO ACCESS A GAME OR CAD APPLICATION UNAVAILABLE ON HIS HOST VALUES PERFORMANCE AND LIKELY FIDELITY. A TECH SUPPORT PERSON MAINTAINING A LIBRARY OF DIFFERENT CONFIGURATIONS AND AN IT ADMINISTRATOR RUNNING SERVER VMS ARE BOTH LIKELY TO VALUE PORTABILITY AND SECURE ISOLATION (INTERPOSITION).

GPU VIRTUALIZATION TAXONOMY• FRONT-END VIRTUALIZATION

- FRONT-END VIRTUALIZATION INTRODUCES A VIRTUALIZATION BOUNDARY AT A RELATIVELY HIGH LEVEL IN THE STACK, AND RUNS THE GRAPHICS DRIVER IN THE HOST/HYPERVISOR. THIS APPROACH DOES NOT RELY ON ANY GPU VENDOR- OR MODEL-SPECIFIC DE- TAILS. • BACK-END VIRTUALIZATION

-THE MOST OBVIOUS BACK-END VIRTUALIZATION TECHNIQUE IS FIXED PASS-THROUGH: THE PERMANENT ASSOCIATION OF A VIRTUAL MACHINE WITH FULL EXCLUSIVE ACCESS TO A PHYSICAL GPU. RECENT CHIPSET FEATURES, SUCH AS INTEL’S VT-D, MAKE FIXED PASS-THROUGH PRACTICAL WITHOUT REQUIRING ANY SPECIAL KNOWLEDGE OF A GPU’S PROGRAMMING INTERFACES. HOWEVER, FIXED PASS-THROUGH IS NOT A GENERAL SOLUTION. IT COMPLETELY FORGOES ANY MULTIPLEXING AND PACKING MACHINES WITH ONE GPU PER VIRTUAL MACHINE (PLUS ONE FOR THE HOST) IS NOT FEASIBLE.

THANKS FOR

WATCHING