Editing Graphics processing unit (section)

=== 2000s ===
NVIDIA released the GeForce 256, marketed as the world's first GPU, integrating transform and lighting engines for advanced 3D graphics rendering. Nvidia was first to produce a chip capable of programmable [[Pixel shader|shading]]: the ''[[GeForce 3]]''. Each pixel could now be processed by a short program that could include additional image textures as inputs, and each geometric vertex could likewise be processed by a short program before it was projected onto the screen. Used in the [[Xbox]] console, this chip competed with the one in the [[PlayStation 2]], which used a custom vector unit for hardware-accelerated vertex processing (commonly referred to as VU0/VU1). The earliest incarnations of shader execution engines used in Xbox were not general-purpose and could not execute arbitrary pixel code. Vertices and pixels were processed by different units, which had their resources, with pixel shaders having tighter constraints (because they execute at higher frequencies than vertices). Pixel shading engines were more akin to a highly customizable function block and did not "run" a program. Many of these disparities between vertex and pixel shading were not addressed until the [[Unified Shader Model]].

In October 2002, with the introduction of the [[ATI Technologies|ATI]] ''[[Radeon 9700 core|Radeon 9700]]'' (also known as R300), the world's first [[Direct3D]] 9.0 accelerator, pixel and vertex shaders could implement [[Loop (computing)|looping]] and lengthy [[floating point]] math, and were quickly becoming as flexible as CPUs, yet orders of magnitude faster for image-array operations. Pixel shading is often used for [[bump mapping]], which adds texture to make an object look shiny, dull, rough, or even round or extruded.<ref>{{cite web | url = https://www.blacksmith-studios.dk/projects/downloads/bumpmapping_using_cg.php | title = Bump Mapping Using CG (3rd Edition) | first = Søren | last = Dreijer | access-date = 2007-05-30 | url-status = dead | archive-url = https://web.archive.org/web/20100120195901/http://www.blacksmith-studios.dk/projects/downloads/bumpmapping_using_cg.php | archive-date = 2010-01-20 }}</ref>

With the introduction of the Nvidia [[GeForce 8 series]] and new generic stream processing units, GPUs became more generalized computing devices. [[Parallel computing|Parallel]] GPUs are making computational inroads against the CPU, and a subfield of research, dubbed GPU computing or [[GPGPU]] for ''general purpose computing on GPU'', has found applications in fields as diverse as [[machine learning]],<ref>{{cite book|chapter=Large-scale deep unsupervised learning using graphics processors |doi=10.1145/1553374.1553486 |publisher=Dl.acm.org |date=2009-06-14 |title=Proceedings of the 26th Annual International Conference on Machine Learning – ICML '09 |pages=1–8 |last1=Raina |first1=Rajat |last2=Madhavan |first2=Anand |last3=Ng |first3=Andrew Y. |s2cid=392458 |isbn=9781605585161 }}</ref> [[oil exploration]], scientific [[image processing]], [[linear algebra]],<ref>[https://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.94.1988&rep=rep1&type=pdf "Linear algebra operators for GPU implementation of numerical algorithms"], Kruger and Westermann, International Conference on Computer Graphics and Interactive Techniques, 2005</ref> [[statistics]],<ref>{{cite journal |title=ABC-SysBio—approximate Bayesian computation in Python with GPU support |last=Liepe |display-authors=etal |journal=Bioinformatics |year=2010 |volume=26 |issue=14 |pages=1797–1799  |doi=10.1093/bioinformatics/btq278 |pmid=20591907 |pmc=2894518 |url=https://bioinformatics.oxfordjournals.org/content/26/14/1797.full |access-date=2010-10-15 |url-status=dead |archive-url=https://web.archive.org/web/20151105144736/https://bioinformatics.oxfordjournals.org/content/26/14/1797.full |archive-date=2015-11-05 }}</ref> [[3D reconstruction]], and [[stock options]] pricing. [[GPGPU]] was the precursor to what is now called a compute shader (e.g. CUDA, OpenCL, DirectCompute) and actually abused the hardware to a degree by treating the data passed to algorithms as texture maps and executing algorithms by drawing a triangle or quad with an appropriate pixel shader.{{clarify|date=April 2023}} This entails some overheads since units like the [[Rasterization|scan converter]] are involved where they are not needed (nor are triangle manipulations even a concern—except to invoke the pixel shader).{{clarify|date=April 2023}}

Nvidia's [[CUDA]] platform, first introduced in 2007,<ref>{{Cite book|url=https://books.google.com/books?id=49OmnOmTEtQC|title=CUDA by Example: An Introduction to General-Purpose GPU Programming, Portable Documents|last1=Sanders|first1=Jason|last2=Kandrot|first2=Edward|date=2010-07-19|publisher=Addison-Wesley Professional|isbn=9780132180139|language=en|url-status=live|archive-url=https://web.archive.org/web/20170412034641/https://books.google.com/books?id=49OmnOmTEtQC|archive-date=2017-04-12}}</ref> was the earliest widely adopted programming model for GPU computing. [[OpenCL]] is an open standard defined by the [[Khronos Group]] that allows for the development of code for both GPUs and CPUs with an emphasis on portability.<ref>{{cite web|url=https://www.khronos.org/opencl/|title=OpenCL – The open standard for parallel programming of heterogeneous systems|work=khronos.org|url-status=live|archive-url=https://web.archive.org/web/20110809103233/https://www.khronos.org/opencl/|archive-date=2011-08-09}}</ref> OpenCL solutions are supported by Intel, AMD, Nvidia, and ARM, and according to a report in 2011 by Evans Data, OpenCL had become the second most popular HPC tool.<ref>{{Cite web |last=Handy |first=Alex |date=2011-09-28 |title=AMD helps OpenCL gain ground in HPC space |url=https://sdtimes.com/amd/amd-helps-opencl-gain-ground-in-hpc-space/ |access-date=2023-06-04 |website=SD Times |language=en-US}}</ref>