GPUs, CPUs and SoC
Download as PPTX, PDF1 like1,535 views
The document discusses the differences between GPUs from NVIDIA and AMD, including architectural comparisons, performance enhancements through multi-GPU setups like CrossFire and SLI, and advanced technologies like NVIDIA PhysX and CUDA. It also touches on the role of integrated vs. dedicated graphics solutions, market share dynamics, and the impact of upcoming technologies like AMD's Mantle and NVIDIA's Tegra. Ultimately, it concludes that while each company has its strengths, competition drives innovation across the industry.
GPUs, CPUs and SoC
- 1. GPUS ON SOC NVIDIA VS. AMD Prepared By: Omar Elshal 19-8014
- 2. OUTLINE • Introduction • CPU vs. GPU & APU • NVidia vs. AMD & INTEL • AMD’s Cross fire vs. NVidia's SLI • NVidia PhysX & CUDA • NVidia Tesla & Tegra • Market share • Who wins ? • References
- 3. INTRODUCTION • Graphics Processing Unit (GPU) is a specialized processor that offloads 3D graphics rendering from the microprocessor. • Used in ES, mobile phones, laptops, tablets and game consoles..etc. • GPU cores are smaller and much more than CPU cores. • CPU have weakness in Linear data execution used for rendering pixels.
- 4. CPU VS. GPU
- 5. APU ARCHITECTURE • Accelerated Processing Unit (APU) combines both CPU and GPU’s features. • Term first introduced by AMD.
- 6. NVIDIA VS. AMD & INTEL • Dedicated GPUs are much more powerful than integrated GPUs. • Most if not all Intel GPUs are integrated ones. • NVidia CUDA cores vs. AMD stream processors. • NVidia's CUDA for parallel programming • AMD powers Xbox one, PS 4 and Wii U by a custom-built APU.
- 7. AMD’S CROSS FIRE VS. NVIDIA'S SLI • Both use the same concept of making use of more than 1 GPU to provide increased graphics performance. • Hybrid SLI enables motherboard GPU to work with Dedicated GPU using GeForce Boost Mode. • Each GPU render different frames allowing optimal performance. • CrossFireX is not limited to using the same GPU model unlike SLI. • HybridPower mode/OPTIMUS allows switching off dedicated GPU if not needed lowering system power, noise and heat.
- 8. NVIDIA ADVANCED PHYSX & 3D VISION • NVIDIA® PhysX® is a powerful physics engine enabling real-physics in video games made exclusively for NVidia GPUs. • PhysX is all about how objects move, interact and react. • CPU does the PhysX for AMD but only NVidia GPUs allows using Advanced PhysX allowing a better visual experience. • 3D Vision kit and glasses allows 3D gaming experience for supported games.
- 9. NVIDIA TESLA & NVIDIA TEGRA • Tesla is used in some of the world’s fastest supercomputers nowadays. • 2008, Tegra is a System on Chip (SoC) series for smartphones, tablets and mobile gaming consoles. • Tegra K1 in Q2 2014.
- 10. TEGRA K1 • The Power Of An Xbox In A Mobile SoC, includes 192 GPU cores • Support pc gaming technologies DirectX 11, OpenGL 4.4 • The first mobile processor to run Unreal Engine 4 • The only mobile GPU to support NVidia CUDA and other NVidia Technologies.
- 11. NVIDIA SHIELD • Uses Tegra 4
- 12. AMD’S MANTLE • AMD’s Mantle is a graphics API like DirectX and OpenGL but provides higher performance. • Utilize the CPU and GPU more efficiently. • AMD claims to be an open API. • Works only on AMD GPUs.
- 13. MARKET SHARE • Intel owns nearly 49.4%, due to their integrated graphics solution being counted. • Not counting those numbers, NVidia and AMD control nearly 100% as of 2008.
- 14. WHO WINS? • For AMD and NVidia, success factors do not differ much. • Every company whether Intel, AMD or NVidia leads in some particular field. • They do not limit their innovation in one field. • All companies started going SoC
- 15. REFERENCES • http://www.nvidia.com/object/tegra-k1.html • Video: http://www.nvidia.com/object/nvision08_gpu_v_cpu.html • Slideshare slides: http://www.slideshare.net/iamsrijon/introduction-to-hybrid-sli-technology • http://www.slideshare.net/CanOzdoruk/nvidia-tesla-k40-gpu • http://www.nvidia.com/object/nvidia-kepler.html • http://www.tomshardware.com/reviews/tegra-k1-kepler-project-denver,3718.html • https://en.wikipedia.org/wiki/AMD • http://www.amd.com/us/products/technologies/mantle/Pages/mantle.aspx#1 • http://www.nvidia.co.uk/object/cuda-parallel-computing-uk.html
Editor's Notes
- #4: Small cores used for executing data simultaneously, Polygon transformation(Floating point type) which are heavily used in Image Processing.
- #5: From 32 cores to 2880 cores in NVidia Tesla K40. floating point operations per second
- #7: In 2006, AMD acquired ATI CUDA (Compute Unified Device Arch.) CUDA™ is a parallel computing platform and programming model that enables dramatic increases in computing performance
- #8: When GeForce Boost mode enabled both GPUs share the rendering load by rendering different frames.
- #9: More realistic experience, walls can be torn down, glass can be shattered, trees bend in the wind, and water flows with body, snow accumulation \ AMD HD3D is 2 years late than NVidia 3D
- #10: In 2012, Audi selected Tegra 3 for its in-vehicle infotainment systems and digital instruments display.
- #11: Tegra K1 uses 64-bit processor with dual-core Project Denver;[56] the world's first 64-bit mobile chip running Android. Project Denver is the codename of a CPU/GPU-microarchitecture designed by Nvidia and based on the ARM instruction set.
- #12: July 2013, NVidia shield firstly presented uses Tegra 4. AMD intends to release ARM System on Chips (SoC) that will begin sampling in early 2014
- #13: However, as of February 2014, the specification has not been made public.
- #15: AMD invests in CPUs and GPUs, they make custom APUs for the three major leading game consoles Xbox, PS, Wii NVidia make their own graphics stuff: CUDA, Tegra (SoC), Shield