?GPU English full name Graphic Processing Unit, Chinese translation for "graphics processor".GPU is a concept relative to the CPU, due to the modern computer (especially home systems, game enthusiasts) graphics processing has become more and more important, the need for a specialized graphics core processor.
The GPU is the "brain" of the graphics card, which determines the grade and most performance of the card, and is also the basis for the difference between 2D and 3D graphics cards. 2D graphics chips rely heavily on the CPU's processing power to handle 3D images and effects. 2D graphics cards rely on the CPU for processing 3D images and special effects, which is called "soft acceleration", while 3D graphics cards centralize the processing of 3D images and special effects within the graphics chip, which is called "hard acceleration". The display chip is usually the largest chip (and the one with the most pins) on a graphics card. Most graphics cards on the market today utilize graphics processing chips from NVIDIA and AMD-ATI.
Today, GPUs are no longer limited to 3D graphics processing, and the development of GPU general-purpose computing technology has attracted a lot of attention in the industry. Facts have also proved that GPUs can provide tens of times or even hundreds of times the performance of a CPU in floating-point computing, parallel computing, and other part of the calculation. "Inevitably, this will make CPU manufacturer Intel nervous for the future, and NVIDIA and Intel often start a war of words over who is more important for the CPU and GPU.The standards for general-purpose computation of GPUs currently include OPEN CL, CUDA, and ATI STREAM. OpenCL (Open Computing Language) is the first open, free standard for parallel programming for general-purpose heterogeneous systems, and a unified programming environment for software developers to write efficient and lightweight code for high-performance computing servers, desktop computing systems, and handheld devices, and is widely used in multi-core processors (CPUs), graphics processing units (GPUs), cell-type architectures, and digital signal processors. type architecture and other parallel processors such as Digital Signal Processors (DSPs), which have broad development prospects in various fields such as gaming, entertainment, scientific research, and medical care, and AMD-ATI and NVIDIA's products now support OPEN CL.
August 20, 1985 ATi Inc. was founded, and in October of the same year, ATi developed the first graphics chip and graphics card using ASIC technology, and the first graphics card using ASIC technology, and the first graphics card using ASIC technology. In April 1992, ATi released the Mach32 graphics card with integrated graphics acceleration. In April 1998, ATi was selected by IDC as the market leader in the graphics chip industry, but at that time there was no GPU designation for this type of chip, and for a long period of time ATI referred to the graphics processor as the VPU until after AMD acquired ATI. It was only after AMD acquired ATI that its graphics chips officially adopted the GPU name.
NVIDIA first introduced the GPU concept in 1999 when it released the GeForce 256 graphics chip. Since then, the core of NVIDIA graphics cards has been referred to by the new name GPU, which allows graphics cards to be less dependent on the CPU and perform some of the work that would otherwise be done by the CPU, especially when it comes to 3D graphics processing. The core technologies used by GPUs include hardware T&L, cubic environment material mapping and vertex blending, texture compression and bump mapping, dual-texture quad-pixel 256-bit rendering engine, and hardware T&L. The GPU is also used to provide a more efficient and reliable way of rendering data.
Major brands
GPUs are produced by a very large number of vendors, and like CPUs, are produced by a large number of vendors, but there are only 3 that are familiar, so much so that people think that there are only 3 vendors of GPUs, AMD, NVIDIA, and Intel.
intel
It may not have occurred to anyone that intel is not only the world's largest producer and seller of CPUs, but also the world's largest producer and seller of GPUs.
intel's GPUs are now fully integrated graphics cards, used in intel motherboards and intel notebooks. In case you didn't realize it, the integrated GPUs that intel ships with its motherboards make up more than 60% of the entire GPU market if you just count the number of units shipped.
Nvidia
is now the largest producer and seller of discrete graphics cards, including the now-familiar
Geforce series, such as the GTX580 GTX480 GTX470 GTX460 GTX295 and so on, as well as the professional Quadro series, and so on.
He also similarly sells integrated graphics cards cured on motherboards, which ship with the motherboard, but NV motherboards have lost half of the graphics card world due to the increased capacity of their own motherboard chips after AMD merged with ATI.
AMD
The world's second largest producer and marketer of discrete graphics cards, formerly known as ATI, produces the familiar HD series, such as the HD7970, HD7950, HD6970, HD6990, and so on. In the professional field there is the FireGL series
As AMD acquired ATI, its motherboard market has expanded comprehensively, and has captured NVIDIA's entire share of AMD processor motherboards. In terms of launch volume and launch profitability, AMD still loses out slightly to NVIDIA in terms of graphics cards, but they are neck and neck, with the difference being just a few percentage points.
MatroxA company that competed with NVIDIA for market share in discrete graphics cards, Matrox's cards and NVIDIA's cards used to be on par in terms of performance at one point in time. However, due to its declining development capabilities, Matrox was gradually phased out of the civilian discrete graphics market during the GF5 period, which was the period of ati's 9000 series, due to its performance lagging behind that of the GF5900 and Raden9800 by an entire generation.
But Matrox still has its place in the world of professional graphics cards for engineering. These cards are still very powerful for engineering mainframes and multitasking. Unlike nv and amd's professional graphics cards, nv and ati's professional graphics cards are in 3D, while Matrox's professional graphics cards are in 2D, which is CAD.
But due to the increasing popularity of cuda, DX10+ cards will be able to perform amazingly well in all CUDA-enabled programs, which means that the day CUDA becomes popular in all sorts of software, Matrox will have a huge impact on the performance of its products.
Matrox will be out of the 2D professional card market the day CUDA becomes commonplace in all kinds of software.
sis and viaSelenium and VIA are now twins, but they used to be two separate companies and both produced their own motherboards with integrated graphics. But the poor brothers have gradually been fading out of the motherboard market, and will surely be fading out of the GPU market as well. With so much at stake for GPUs, perhaps a new manufacturer will be born in the future as well.
Other informationNVIDIA first introduced the concept of GPUs in 1999 with the release of the GeForce256 graphics chip. GPUs enable graphics cards to reduce their dependence on the CPU and do some of the work that the CPU would otherwise do, especially when it comes to 3D graphics processing. The core technologies used by GPUs are hard T&L, Cubic Environment Material Mapping and Vertex Blending, Texture Compression, and Vertex Blending. Vertex blending, texture compression and bump mapping, dual-texture quad-pixel 256-bit rendering engine, etc., and the hard T&L technology can be said to be the hallmark of GPUs. GPU is the "heart" of the graphics card, which is equivalent to the role of the CPU in a computer, which determines the grade and most of the performance of the graphics card, and is also the basis for the difference between 2D and 3D graphics cards. 2D graphics cards are more efficient in processing 3D images and special effects. The display chip is usually the largest chip (and the one with the most pins) on a graphics card. Most graphics cards on the market today utilize graphics processing chips from NVIDIA and AMD.
GPUs are able to support T&L (TransformandLighting) display chips on their hardware because T&L is an important part of 3D rendering, which calculates the 3D position of polygons and handles dynamic lighting effects, also known as "geometry processing". geometry processing". A good T&L unit can provide detailed 3D objects and advanced lighting effects; except that in most PCs, most of the T&L operations are left to the CPU (which is also known as soft T&L), due to the CPU's many tasks, in addition to T&L, but also to do the memory management, input echo, and other non-3D graphics processing Work, so in the actual computing time performance will be greatly reduced, often appearing in the display card waiting for the CPU information, its computing speed is far from keeping up with today's complex three-dimensional game requirements. Even if the CPU's operating frequency is above 1GHz or higher, it will not help much, as this is a problem caused by the design of the PC itself and has little to do with the speed of the CPU.
How it worksSimply put, a GPU is a graphics chip that supports T&L (Transform and Lighting) on a hardware level, as T&L is an important part of 3D rendering that calculates the 3D position of polygons and processes them. Its role is to calculate the 3D position of polygons and handle dynamic lighting effects, also known as "geometry processing". A good T&L unit can provide detailed 3D objects and advanced lighting effects; however, in most PCs, most of the T&L operations are handled by the CPU (which is also known as soft T&L), and because the CPU has many tasks, in addition to T&L, it also has to do memory management, input echo, and other non-3D graphics processing work. Work, so in the actual computing time performance will be greatly reduced, often appearing in the display card waiting for the CPU information, its computing speed is far from keeping up with today's complex three-dimensional game requirements. Even if the CPU's operating frequency is more than 1GHz or higher, it doesn't help much, because it's a problem caused by the design of the PC itself, and has little to do with the speed of the CPU.
Differences from DSPThe GPU differs from the DSP (Digital Signal Processing) architecture in several key ways. All calculations use floating-point algorithms, and there are currently no bit or integer arithmetic instructions. In addition, since the GPU is designed for image processing, the storage system is actually a two-dimensional segmented storage space consisting of a segment number (from which the image is read) and a two-dimensional address (the X and Y coordinates in the image). In addition, there are no indirect write instructions. The output write address is determined by the raster processor and cannot be changed by the program. This is a great challenge for algorithms that are naturally distributed in memory. Finally, no communication is allowed between the processing of different fragments. In fact, the fragment processor is a SIMD data parallel execution unit that executes code independently in all fragments.
Despite these constraints, GPUs can efficiently perform a wide range of operations, from linear algebra and signal processing to numerical simulation. Despite the simplicity of the concept, new users can be confused when computing with GPUs, which require proprietary graphics knowledge. In this case, some softwares tools can help. Two high-level shading languages, CG and HLSL, allow users to write C-like code that is later compiled into fragmented assembly language. brook is a high-level language designed for GPU computing that does not require graphics knowledge. Therefore, it can be a good starting point for first-time developers working with GPUs. Brook is an extension of the C language, incorporating simple data parallel programming constructs that can be directly translated to GPUs. The data stored and manipulated by the GPU is visualized as a stream, similar to an array in standard C. The kernel is a stream of data that is stored in the GPU and manipulated by the GPU. A kernel is a function that operates on a stream. Calling a kernel function on a series of input streams implies an implicit loopback on the stream elements, i.e., calling the kernel body for each stream element. brook also provides simplification mechanisms, such as summing, maximizing, or multiplying all the elements of a stream. brook also completely hides all the details of the graphical APIs and virtualizes parts of the GPU that are unfamiliar to many users, such as the two-dimensional memory system. unfamiliar parts of the GPU, such as the two-dimensional memory system, were virtualized. Applications written in Brook include linear algebra subroutines, fast Fourier transforms, ray tracing, and image processing. Utilizing ATI's X800XT and Nvidia's GeForce 6800 Ultra GPUs, many of these routines are up to seven times faster with the same high-speed caching, SSE assembly-optimized Pentium 4 execution.
Users interested in GPU computing have struggled to adapt algorithms to graphical fundamentals. The availability of high-level programming languages like Brook makes it easy for even novice programmers to grasp the performance benefits of the GPU. The ease of access to GPU computing capabilities also allows the evolution of the GPU to continue, not just as a graphics engine, but as the primary computing engine for PCs.
Identification softwaresWhen it comes to processor identification tools, CPU-Z needs no introduction. Hardware site TechPowerUp gives us a similar tool for graphics card identification, "GPU-Z", which supports XP/Vista/2000/2003. The latest version is 0.6.2, updated on May 2, 2012, with a file size of 1MB.
The latest version is 0.6.2, updated on May 2, 2012, with a file size of 1MB.
The latest version is 0.6.2, updated on May 2, 2012, with a file size of 1MB. Update Notes:
GPU-Z 0.6.2
Fixed NVIDIA Kepler Boost operating frequency display issue
GPU-Z 0.6.1
Added support for GeForce605, NVIDIA GeForce GTX690, GTX 670, GT 640, GT 630
Added support for AMD Radeon HD7970M, HD7450
Added support for Intel Sandy Bridge and Ivy Bridge integrated graphics cards
Fixed GK107 ROP detection and display
Improved support for NVIDIA Kelvin and NVIDIA Kelvin. Improved detection of NVIDIA Kepler Boost operating frequency
Fixed a GPU-Z crash issue on AMD Southern Islands graphics cards when no driver is installed
Fixed an issue that does not work correctly on Windows XP
Disabled NVIDIA Kepler Incomplete BIOS Storage
Disabled NVIDIA Kepler Incomplete BIOS Storage
Disabled NVIDIA Kepler Boosting.
BIOS files are now stored separately as .rom
Fixed sensor graph redrawing and improved performance
Related IssuesFirst issue:
GPU competition is far more intense than CPU competition. There are just two major manufacturers of CPUs for general purpose PCs, Intel and AMD. The leading GPU manufacturers are NVIDIA and AMD, but there are also Intel, Samsung, and several other manufacturers that can produce low-end and mid-range products. Although their products are not as good as the first two, but in many applications can also meet the needs of users, so N and A record only desperately run forward in order not to die. CPU manufacturers did not adopt the advanced technology of the GPU is because CPU manufacturers have their own investment in the production line, it is not possible to put the original production line are eliminated on the new production line, which may be difficult to recover the funds invested in the first place. GPU manufacturers for various reasons, are generally their own design by others OEM, such as finding TSMC OEM. OEM manufacturers in order to receive business, only to keep upgrading their production equipment, so as to survive. That's why it's so important to have the above.
The second problem
The CPU handles the AI, plot, and other aspects of the game, as well as some of the graphical aspects. When Microsoft releases a new DX each time, not every GPU can support the new DX features, so some of the graphics aspects are still done by the CPU. There are also some features like gravity features that used to be done by the CPU, but now that some GPUs can support them, those tasks are done by the GPU.
The third problem
The GPU is equivalent to a CPU dedicated to image processing, and because it is specialized, it is strong, and it works much more efficiently than the CPU when processing images, but the CPU is a general-purpose data processor, and it is its strongest when dealing with numerical calculations, and it is not able to replace the tasks it can accomplish with the GPU, so it cannot be replaced by the GPU.
In addition, there are some features, such as gravity features, that were previously done by the CPU, and now some of them are also supported by the GPU.
In addition
AMD acquired ATI for $5.4 billion in 2006, and AMD has seen that CPUs and GPUs can only take the path of convergence in order to take advantage of the competition in the future. how to work with CPUs and GPUs to maximize the efficiency of the work is a problem that AMD is now considering, and it is also a problem that Intel is facing.
The fourth issue
Microsoft released Windows 7, one of the notable features of the joint GPU and CPU power, to enhance the value of the GPU in the use of hardware, in Windows 7, the CPU and the GPU form a co-processing environment. the CPU calculates a very complex sequence of code, while the GPU is running massively parallel applications. Microsoft utilizes DirectX Compute to make the GPU one of the core components of the operating system. directX Compute. it allows developers to leverage the massively parallel computing power of the GPU to create compelling consumer and professional computing applications. Simply put, DirectX Compute is the GPU general purpose computing interface developed by Microsoft to unify the GPU general purpose computing standard. In other words, after Windows 7, GPUs will be second only to CPUs in terms of hardware status and will be able to play a greater role. Intel's Core 2.0 Sandy Bridge processors with integrated core graphics and AMD Llano APUs have elevated integrated graphics to a new level, and Intel's next-generation Ivy Bridge processors have further enhanced the capabilities of core graphics.