GeForce 900 series

{{Short description|Series of GPUs by Nvidia}} {{Redirect|GM200|text=For the radars, see Ground Master 200 and Ground Master 200 Multi Mission.}} {{for|GeForce cards with a model number of 9XX0|GeForce 9 series}} {{Use mdy dates|date=February 2016}} {{Infobox GPU | name = GeForce 900 series | image = 300px 300px | caption = '''Top''': Logo of the series '''Bottom''': A GeForce GTX 980 Ti released in 2015, the series' highest-end non-Titan model | codename = GM20x | architecture = Maxwell | created = {{start date and age|September 18, 2014}} | model = GeForce series | model1 = GeForce GT series | model2 = GeForce GTX series | transistors = 2.94B (GM206) | transistors1 = 5.2B (GM204) | transistors2 = 8.0B (GM200) | transistors3 = | transistors4 = | transistors5 = | transistors6 = | fab = TSMC | process = 28 nm | entry = | midrange = {{Unbulleted list|GeForce GTX 950|GeForce GTX 960}} | highend = {{Unbulleted list|GeForce GTX 970|GeForce GTX 980}} | enthusiast = {{Unbulleted list|GeForce GTX 980 Ti|Nvidia Titan X}} | openglversion = OpenGL 4.6 | d3dversion = Direct3D 12 (feature level 12_1)<ref name="D3D11.3">{{cite web|url=http://www.anandtech.com/show/8526/nvidia-geforce-gtx-980-review/4|archive-url=https://web.archive.org/web/20140922001359/http://www.anandtech.com/show/8526/nvidia-geforce-gtx-980-review/4|url-status=dead|archive-date=September 22, 2014|title=Maxwell 2's New Features: Direct3D 11.3 & VXGI - The NVIDIA GeForce GTX 980 Review: Maxwell Mark 2|author=Ryan Smith|work=anandtech.com}}</ref><ref name="dx12_1">{{cite web|url=http://blogs.nvidia.com/blog/2014/09/19/maxwell-and-dx12-delivered/|title=Maxwell and DirectX 12 Delivered|work=The Official NVIDIA Blog}}</ref><ref>{{cite web|url=http://blogs.msdn.com/b/directx/archive/2014/09/18/directx-12-lights-up-nvidia-s-maxwell-editor-s-day.aspx|title=MSDN Blogs|publisher=Microsoft|work=msdn.com}}</ref><ref name="11.3And12.0RenderingFeaturesAreEquivalent">{{cite web|url=http://www.anandtech.com/show/8544/microsoft-details-direct3d-113-12-new-features|archive-url=https://web.archive.org/web/20140921000530/http://www.anandtech.com/show/8544/microsoft-details-direct3d-113-12-new-features|url-status=dead|archive-date=September 21, 2014|title=Microsoft Details Direct3D 11.3 & 12 New Rendering Features|author=Ryan Smith|work=anandtech.com}}</ref> Shader Model 6.7 | openclversion = OpenCL 3.0{{efn|name="OpenCL3.0"|In OpenCL 3.0, OpenCL 1.2 functionality has become a mandatory baseline, while all OpenCL 2.x and OpenCL 3.0 features were made optional.}} | vulkanapi = Vulkan 1.4<ref name="vulkan">{{cite web |url=https://developer.nvidia.com/vulkan-driver |title=Vulkan Driver Support |work=Nvidia |date=February 10, 2016 |access-date=2018-04-25}}</ref> SPIR-V | predecessor = {{Unbulleted list|GeForce 800M series|GeForce 700 series}} | successor = GeForce 10 series | support status = Game-ready Driver support ended in November 2025.<ref>{{Cite web |date=2025-11-04 |url=https://www.nvidia.com/en-us/drivers/details/257496/|title=GeForce Game Ready Driver 581.80|access-date=2026-02-19 |work=Nvidia}}</ref> Security updates until October 2028.<ref>{{Cite web |last=Kampman |first=Jeffrey |date=2025-07-31 |title=Nvidia confirms end of Game Ready driver support for Maxwell and Pascal GPUs — affected products will get optimized drivers through October 2025 |url=https://www.tomshardware.com/pc-components/gpus/nvidia-confirms-end-of-game-ready-driver-support-for-maxwell-and-pascal-gpus-affected-products-will-get-optimized-drivers-through-october-2025 |access-date=2025-08-21 |website=Tom's Hardware |language=en}}</ref> }} The '''GeForce 900 series''' is a family of graphics processing units developed by Nvidia, succeeding the GeForce 700 series and serving as the high-end introduction to the Maxwell microarchitecture, named after James Clerk Maxwell. They were produced with TSMC's 28 nm process.

With Maxwell, the successor to Kepler, Nvidia expected three major outcomes: improved graphics capabilities, simplified programming, and better energy efficiency compared to the GeForce 700 series and GeForce 600 series.<ref name=Xbitlabs>{{cite web|url=http://www.xbitlabs.com/news/graphics/display/20130412175120_Nvidia_Next_Generation_Maxwell_Architecture_Will_Break_New_Grounds.html|title=Nvidia: Next-Generation Maxwell Architecture Will Break New Grounds - X-bit labs|work=xbitlabs.com|url-status=dead|archive-url=https://web.archive.org/web/20130629025423/http://www.xbitlabs.com/news/graphics/display/20130412175120_Nvidia_Next_Generation_Maxwell_Architecture_Will_Break_New_Grounds.html|archive-date=June 29, 2013|df=mdy-all}}</ref>

Maxwell was announced in September 2010,<ref>{{cite web|url=http://www.anandtech.com/show/3939/gtc-2010-reporters-notebook-day-1-nvidia-announces-future-gpu-families-for-2011-and-2013|archive-url=https://web.archive.org/web/20100924065622/http://www.anandtech.com/show/3939/gtc-2010-reporters-notebook-day-1-nvidia-announces-future-gpu-families-for-2011-and-2013|url-status=dead|archive-date=September 24, 2010|title=GTC 2010 Day 1: NVIDIA Announces Future GPU Families for 2011 And 2013|author=Ryan Smith|work=anandtech.com}}</ref> with the first Maxwell-based GeForce consumer-class products released in early 2014.<ref>{{cite web|url=https://www.geforce.com/whats-new/articles/introducing-the-geforce-gtx-750-class|title=GeForce GTX 750 Class GPUs: Serious Gaming, Incredible Value|work=geforce.com}}</ref>

== Architecture == {{main|Maxwell (microarchitecture)}}

=== {{Anchor|SMM}}First generation Maxwell (GM10x) === First generation Maxwell GM107/GM108 were released as GeForce GTX 745, GTX 750/750 Ti and GTX 850M/860M (GM107) and GT 830M/840M (GM108). These new chips provide few consumer-facing additional features; Nvidia instead focused on power efficiency. Nvidia increased the amount of L2 cache from 256 KiB on GK107 to 2 MiB on GM107, reducing the memory bandwidth needed. Accordingly, Nvidia cut the memory bus from 192 bit on GK106 to 128 bit on GM107, further saving power.<ref name="anand750"/> Nvidia also changed the streaming multiprocessor design from that of Kepler (SMX), naming it SMM. The structure of the warp scheduler is inherited from Kepler, which allows each scheduler to issue up to two instructions that are independent from each other and are in order from the same warp. The layout of SMM units is partitioned so that each of the 4 warp schedulers in an SMM controls 1 set of 32 FP32 CUDA cores, 1 set of 8 load/store units, and 1 set of 8 special function units. This is in contrast to Kepler, where each SMX has 4 schedulers that schedule to a shared pool of 6 sets of 32 FP32 CUDA cores, 2 sets of 16 load/store units, and 2 sets of 16 special function units.<ref name="AnandTechGTX750">{{cite web|url=http://www.anandtech.com/show/7764/the-nvidia-geforce-gtx-750-ti-and-gtx-750-review-maxwell/3|archive-url=https://web.archive.org/web/20140221105651/http://anandtech.com/show/7764/the-nvidia-geforce-gtx-750-ti-and-gtx-750-review-maxwell/3|url-status=dead|archive-date=February 21, 2014|title=Maxwell: Designed For Energy Efficiency - The NVIDIA GeForce GTX 750 Ti and GTX 750 Review: Maxwell Makes Its Move|author=Ryan Smith, Ganesh T S|work=anandtech.com}}</ref> These units are connected by a crossbar that uses power to allow the resources to be shared.<ref name="AnandTechGTX750"/> This crossbar is removed in Maxwell.<ref name="AnandTechGTX750"/> Texture units and FP64 CUDA cores are still shared.<ref name="anand750"/> SMM allows for a finer-grain allocation of resources than SMX, saving power when the workload isn't optimal for shared resources. Nvidia claims a 128 CUDA core SMM has 86% of the performance of a 192 CUDA core SMX.<ref name="anand750"/> Also, each Graphics Processing Cluster, or GPC, contains up to 4 SMX units in Kepler, and up to 5 SMM units in first generation Maxwell.<ref name="anand750"/>

GM107 supports CUDA Compute Capability 5.0 compared to 3.5 on GK110/GK208 GPUs and 3.0 on GK10x GPUs. Dynamic Parallelism and HyperQ, two features in GK110/GK208 GPUs, are also supported across the entire Maxwell product line.

Maxwell provides native shared memory atomic operations for 32-bit integers and native shared memory 32-bit and 64-bit compare-and-swap (CAS), which can be used to implement other atomic functions.

While it was once thought that Maxwell used tile-based immediate mode rasterization,<ref>{{Cite news | url = https://www.realworldtech.com/tile-based-rasterization-nvidia-gpus/ | title = Tile-based Rasterization in Nvidia GPUs | first = David | last = Kanter | date = August 1, 2016 | newspaper = Real World Technologies | access-date = August 16, 2016}}</ref> Nvidia corrected this at GDC 2017 saying Maxwell instead uses Tile Caching.<ref>{{Cite news | url = https://www.hardware.fr/news/15027/gdc-nvidia-parle-tile-caching-maxwell-pascal.html | title = GDC: Nvidia talks about Tile Caching by Maxwell and Pascal | first = Damien | last = Triolet | date = March 3, 2017 | newspaper = Hardware.fr | access-date = May 24, 2017}}</ref>

==== NVENC ==== {{Main|Nvidia NVENC}} Maxwell-based GPUs also contain the NVENC SIP block introduced with Kepler. Nvidia's video encoder, NVENC, is 1.5 to 2 times faster than on Kepler-based GPUs meaning it can encode video at 6 to 8 times playback speed.<ref name="anand750">{{cite news|last1=Smith|first1=Ryan|last2=T S|first2=Ganesh|title=The NVIDIA GeForce GTX 750 Ti and GTX 750 Review: Maxwell Makes Its Move|url=http://www.anandtech.com/show/7764/the-nvidia-geforce-gtx-750-ti-and-gtx-750-review-maxwell|publisher=AnandTech|access-date=February 18, 2014|archive-url=https://web.archive.org/web/20140218145134/http://www.anandtech.com/show/7764/the-nvidia-geforce-gtx-750-ti-and-gtx-750-review-maxwell|archive-date=February 18, 2014|date=February 18, 2014}}</ref>

==== PureVideo ==== {{Main|Nvidia PureVideo}} Nvidia also claims an 8 to 10 times performance increase in PureVideo Feature Set E video decoding due to the video decoder cache paired with increases in memory efficiency. However, H.265 is not supported for full hardware decoding, relying on a mix of hardware and software decoding.<ref name="anand750"/> When decoding video, a new low power state "GC5" is used on Maxwell GPUs to conserve power.<ref name="anand750"/>

=== Second generation Maxwell (GM20x) === Second generation Maxwell introduced several new technologies: Dynamic Super Resolution,<ref>{{cite web|url=https://www.geforce.com/whats-new/articles/dynamic-super-resolution-instantly-improves-your-games-with-4k-quality-graphics|title=Dynamic Super Resolution Improves Your Games With 4K-Quality Graphics On HD Monitors|work=geforce.com}}</ref> Third Generation Delta Color Compression,<ref name="international.download.nvidia.com">{{cite web |url=http://international.download.nvidia.com/geforce-com/international/pdfs/GeForce_GTX_980_Whitepaper_FINAL.PDF |title=Whitepaper: NVIDIA GeForce GTX 980 |access-date=September 20, 2014 |archive-url=https://web.archive.org/web/20170721113746/http://international.download.nvidia.com/geforce-com/international/pdfs/GeForce_GTX_980_Whitepaper_FINAL.PDF |archive-date=July 21, 2017|url-status=dead }}</ref> Multi-Pixel Programming Sampling,<ref>{{cite web|url=http://www.geforce.com/hardware/technology/mfaa/technology%7ctitle=MFAA%7cwork=geforce.com|title=NVIDIA - Maintenance|work=geforce.com}}</ref> Nvidia VXGI (Real-Time-Voxel-Global Illumination),<ref>{{cite web|url=https://www.geforce.com/whats-new/articles/maxwells-voxel-global-illumination-technology-introduces-gamers-to-the-next-generation-of-graphics|title=Maxwell's Voxel Global Illumination Technology Introduces Gamers To The Next Generation Of Graphics|work=geforce.com}}</ref> VR Direct,<ref>{{cite web|url=https://www.geforce.com/whats-new/articles/maxwell-architecture-gpus-the-only-choice-for-virtual-reality-gaming|title=NVIDIA Maxwell GPUs: The Best Graphics Cards For Virtual Reality Gaming|work=geforce.com}}</ref><ref>{{cite web|url=http://blogs.nvidia.com/blog/2014/09/18/maxwell-virtual-reality/|title=How Maxwell's VR Direct Brings Virtual Reality Gaming Closer to Reality|work=The Official NVIDIA Blog}}</ref><ref name="anandtech.com">{{cite web|url=http://www.anandtech.com/show/8526/nvidia-geforce-gtx-980-review/5|archive-url=https://web.archive.org/web/20140922001403/http://www.anandtech.com/show/8526/nvidia-geforce-gtx-980-review/5|url-status=dead|archive-date=September 22, 2014|title=Display Matters: HDMI 2.0, HEVC, & VR Direct - The NVIDIA GeForce GTX 980 Review: Maxwell Mark 2|author=Ryan Smith|work=anandtech.com}}</ref> Multi-Projection Acceleration,<ref name="international.download.nvidia.com"/> and Multi-Frame Sampled Anti-Aliasing (MFAA)<ref>{{cite web|url=https://www.geforce.com/whats-new/articles/multi-frame-sampled-anti-aliasing-delivers-better-performance-and-superior-image-quality|title=Multi-Frame Sampled Anti-Aliasing Delivers Better Performance To Maxwell Gamers|work=geforce.com}}</ref> (however support for Coverage-Sampling Anti-Aliasing (CSAA) was removed).<ref>{{cite web|url=https://forums.realhardwarereviews.com/topic/260/new-nvidia-maxwell-chips-do-not-support-fast-csaa|title=New nVidia Maxwell chips do not support fast CSAA|work=realhardwarereviews.com|access-date=May 7, 2019|archive-date=May 7, 2019|archive-url=https://web.archive.org/web/20190507040023/https://forums.realhardwarereviews.com/topic/260/new-nvidia-maxwell-chips-do-not-support-fast-csaa|url-status=dead}}</ref> HDMI 2.0 support was also added.<ref>{{cite web|url=https://www.geforce.com/whats-new/articles/maxwell-architecture-gtx-980-970|title=Introducing The Amazing New GeForce GTX 980 & 970|work=geforce.com}}</ref><ref>{{cite web|url=http://www.anandtech.com/show/8526/nvidia-geforce-gtx-980-review|archive-url=https://web.archive.org/web/20140922001256/http://www.anandtech.com/show/8526/nvidia-geforce-gtx-980-review|url-status=dead|archive-date=September 22, 2014|title=The NVIDIA GeForce GTX 980 Review: Maxwell Mark 2|author=Ryan Smith|work=anandtech.com}}</ref>

Second generation Maxwell also changed the ROP to memory controller ratio from 8:1 to 16:1.<ref name="AnandTech980page3">{{cite web|url=http://www.anandtech.com/show/8526/nvidia-geforce-gtx-980-review/3|archive-url=https://web.archive.org/web/20140922001353/http://www.anandtech.com/show/8526/nvidia-geforce-gtx-980-review/3|url-status=dead|archive-date=September 22, 2014|title=Maxwell 2 Architecture: Introducing GM204 - The NVIDIA GeForce GTX 980 Review: Maxwell Mark 2|author=Ryan Smith|work=anandtech.com}}</ref> However, some of the ROPs are generally idle in the GTX 970 because there are not enough enabled SMMs to give them work to do and therefore reduces its maximum fill rate.<ref name="techreport.com">{{cite web|url=http://techreport.com/blog/27143/here-another-reason-the-geforce-gtx-970-is-slower-than-the-gtx-980|title=Here's another reason the GeForce GTX 970 is slower than the GTX 980|work=techreport.com|date=October 2014}}</ref>

Second generation upgraded NVENC which supports HEVC encoding and adds support for H.264 encoding resolutions at 1440p/60FPS & 4K/60FPS compared to NVENC on Maxwell first generation GM10x GPUs which only supported H.264 1080p/60FPS encoding.<ref name="anandtech.com" /> Maxwell GM206 GPU supports full fixed function HEVC hardware decoding.<ref>{{cite web|url=http://www.anandtech.com/show/8923/nvidia-launches-geforce-gtx-960|archive-url=https://web.archive.org/web/20150124021316/http://www.anandtech.com/show/8923/nvidia-launches-geforce-gtx-960|url-status=dead|archive-date=January 24, 2015|title=NVIDIA Launches GeForce GTX 960|author=Ryan Smith|work=anandtech.com}}</ref><ref>{{cite web|url=http://www.anandtech.com/show/9547/nvidia-launches-geforce-gtx-950-gm206-the-lesser-for-159|archive-url=https://web.archive.org/web/20150820230835/http://www.anandtech.com/show/9547/nvidia-launches-geforce-gtx-950-gm206-the-lesser-for-159|url-status=dead|archive-date=August 20, 2015|title=NVIDIA Launches GeForce GTX 950; GM206 The Lesser For $159|author=Ryan Smith|work=anandtech.com}}</ref>

===Async compute support=== While the Maxwell series was marketed as fully DirectX 12 compliant,<ref name="dx12_1"/><ref name="980tidx12">[https://international.download.nvidia.com/geforce-com/international/images/nvidia-geforce-gtx-980-ti/nvidia-geforce-gtx-980-ti-directx-12-advanced-api-support.png Advanced API support] nvidia.com</ref><ref name="auto">{{cite web|url=http://www.geforce.com/hardware/desktop-gpus/geforce-gtx-980/specifications|title=GeForce GTX 980 - Specifications - GeForce|work=geforce.com}}</ref> Oxide Games, developer of ''Ashes of the Singularity'', uncovered that Maxwell-based cards do not perform well when async compute is utilized.<ref>{{cite web|url=http://www.pcper.com/reviews/Graphics-Cards/DX12-GPU-and-CPU-Performance-Tested-Ashes-Singularity-Benchmark/Results-Avera|title=DX12 GPU and CPU Performance Tested: Ashes of the Singularity Benchmark|work=pcper.com|access-date=August 31, 2015|archive-url=https://web.archive.org/web/20160415224825/http://www.pcper.com/reviews/Graphics-Cards/DX12-GPU-and-CPU-Performance-Tested-Ashes-Singularity-Benchmark/Results-Avera|archive-date=April 15, 2016|url-status=dead}}</ref><ref name="oxidedx12">{{cite web|url=https://www.guru3d.com/news-story/nvidia-wanted-oxide-dev-dx12-benchmark-to-disable-certain-settings.html|title=Nvidia Wanted Oxide dev DX12 benchmark to disable certain DX12 Features ? (content updated)|author=Hilbert Hagedoorn|work=Guru3D.com|date=August 31, 2015 }}</ref><ref name="nvidiashim">{{cite web|url=https://www.oxidegames.com/2015/08/16/the-birth-of-a-new-api/|title=The Birth of a new API|date=August 16, 2015|work=Oxide Games}}</ref><ref name="980tidx12" />

It appears that while this core feature is in fact exposed by the driver,<ref>{{cite web| url=https://www.overclock.net/t/1569897/various-ashes-of-the-singularity-dx12-benchmarks/2130#post_24379702| title=[Various] Ashes of the Singularity DX12 Benchmarks| date=August 17, 2015|work=Overclock.net}}</ref> Nvidia partially implemented it through a driver-based shim, coming at a high performance cost.<ref name="nvidiashim" /> Unlike AMD's competing GCN-based graphics cards which include a full implementation of hardware-based asynchronous compute,<ref>{{cite web|url=https://www.techpowerup.com/215663/lack-of-async-compute-on-maxwell-makes-amd-gcn-better-prepared-for-directx-12.html|title=Lack of Async Compute on Maxwell Makes AMD GCN Better Prepared for DirectX 12|work=TechPowerUp|date=August 31, 2015 }}</ref><ref>{{cite web|url=https://www.guru3d.com/articles-pages/amd-radeon-r9-fury-x-review,8.html|title=AMD Radeon R9 Fury X review|author=Hilbert Hagedoorn|work=Guru3D.com|date=June 24, 2015 }}</ref> Nvidia planned to rely on the driver to implement a software queue and a software distributor to forward asynchronous tasks to the hardware schedulers, capable of distributing the workload to the correct units.<ref>{{cite web| url=https://www.overclock.net/t/1569897/various-ashes-of-the-singularity-dx12-benchmarks/2170#post_24380072 | title=[Various] Ashes of the Singularity DX12 Benchmarks|date=August 17, 2015 | work=Overclock.net}}</ref> Asynchronous compute on Maxwell therefore requires that both a game and the GPU driver be specifically coded for asynchronous compute on Maxwell in order to enable this capability.<ref name="PCPerTimeSpy">{{Cite news | first=Ryan | last=Shrout | url=http://www.pcper.com/reviews/Graphics-Cards/3DMark-Time-Spy-Looking-DX12-Asynchronous-Compute-Performance | title=3DMark Time Spy: Looking at DX12 Asynchronous Compute Performance | date=July 14, 2016 | newspaper=PC Perspective | access-date=July 14, 2016 | archive-url=https://web.archive.org/web/20160715155102/http://www.pcper.com/reviews/Graphics-Cards/3DMark-Time-Spy-Looking-DX12-Asynchronous-Compute-Performance | archive-date=July 15, 2016 | url-status=dead }}</ref> The 3DMark Time Spy benchmark shows no noticeable performance difference between asynchronous compute being enabled or disabled.<ref name="PCPerTimeSpy"/> Asynchronous compute is disabled by the driver for Maxwell.<ref name="PCPerTimeSpy"/>

Oxide claims that this led to Nvidia pressuring them not to include the asynchronous compute feature in their benchmark at all, so that the 900 series would not be at a disadvantage against AMD's products which implement asynchronous compute in hardware.<ref name="oxidedx12"/>

Maxwell requires that the GPU be statically partitioned for asynchronous compute to allow tasks to run concurrently.<ref name="RyanSmithDynamicScheduling">{{Cite news | url=http://www.anandtech.com/show/10325/the-nvidia-geforce-gtx-1080-and-1070-founders-edition-review/9 | archive-url=https://web.archive.org/web/20160723082341/http://www.anandtech.com/show/10325/the-nvidia-geforce-gtx-1080-and-1070-founders-edition-review/9 | url-status=dead | archive-date=July 23, 2016 | title=The NVIDIA GeForce GTX 1080 & GTX 1070 Founders Editions Review: Kicking Off the FinFET Generation | first=Ryan | last=Smith | date=July 20, 2016 | access-date=July 21, 2016 | newspaper=AnandTech | page=9}}</ref> Each partition is assigned to a hardware queue. If any of the queues that are assigned to a partition empty out or are unable to submit work for any reason (e.g. a task in the queue must be delayed until a hazard is resolved), the partition and all of the resources in that partition reserved for that queue will idle.<ref name="RyanSmithDynamicScheduling"/> Asynchronous compute therefore could easily hurt performance on Maxwell if it is not coded to work with Maxwell's static scheduler.<ref name="RyanSmithDynamicScheduling"/> Furthermore, graphics tasks saturate Nvidia GPUs much more easily than they do to AMD's GCN-based GPUs which are much more heavily weighted towards compute, so Nvidia GPUs have fewer scheduling holes that could be filled by asynchronous compute than AMD's.<ref name="RyanSmithDynamicScheduling"/> For these reasons, the driver forces a Maxwell GPU to place all tasks into one queue and execute each task in serial, and give each task the undivided resources of the GPU no matter whether or not each task can saturate the GPU or not.<ref name="RyanSmithDynamicScheduling"/>

==Products==

===GeForce 900 (9xx) series=== {{See also|List of Nvidia graphics processing units#GeForce 900 series|l1 = Comparison table of GeForce 900 series}} {{row hover highlight}} {{sort under}} {| class="wikitable sortable sort-under mw-datatable" style="font-size: 80%; text-align: center;" |- ! rowspan="2" | Model ! rowspan="2" | Launch ! rowspan="2" | Code name ! rowspan="2" | Process ! rowspan="2" | Transistors (billion) ! rowspan="2" | Die size (mm2) ! rowspan="2" | Core config{{efn|name=CoreConfig}} ! rowspan="2" | Bus interface ! rowspan="2" |L2 Cache (MB) ! colspan="3" | Clock Speeds ! colspan="4" | Memory ! colspan="2" | Fillrate{{efn|name=PerfValues}} ! colspan="2" | Processing power (GFLOPS){{efn|name=PerfValues}}{{efn|name=ProcessingPower}} ! rowspan="2" | TDP (Watts) ! rowspan="2" | SLI support ! Release price (USD) |- ! Base (MHz) ! Boost (MHz) ! Memory (MT/s) ! Size (GB) ! Bandwidth (GB/s) ! Bus type ! Bus width (bit) ! Pixel (GP/s){{efn|name=PixelFillrate}} ! Texture (GT/s){{efn|name=TextureFillrate}} ! Single precision ! Double precision ! MSRP |- ! style="text-align:left;" | GeForce GT 945A<ref>{{cite web |url=https://support.hp.com/at-de/document/c04996577 |title=Sprout Pro by HP |publisher=HP |access-date=2019-01-09 |archive-url=https://web.archive.org/web/20190109111211/https://support.hp.com/at-de/document/c04996577 |archive-date=2019-01-09 |url-status=live }}</ref><ref>{{cite web |url=https://devtalk.nvidia.com/default/topic/915766/linux-solaris-and-freebsd-driver-361-28-long-lived-branch-release-/ |title=Linux, Solaris, and FreeBSD driver 361.28 (long-lived branch release) |publisher=Nvidia |date=2016-02-09 |access-date=2016-02-10 |archive-url=https://web.archive.org/web/20160216022209/https://devtalk.nvidia.com/default/topic/915766/linux-solaris-and-freebsd-driver-361-28-long-lived-branch-release-/ |archive-date=2016-02-16 |url-status=live }}</ref><ref>{{cite web |url=https://www.techpowerup.com/gpu-specs/geforce-945a.c2813 |title=NVIDIA GeForce 945A Specs |access-date=2018-08-06 |url-status=live |archive-url=https://archive.today/20241103051053/https://www.techpowerup.com/gpu-specs/geforce-945a.c2813 |archive-date=2024-11-03}}</ref> | February, 2016 | GM108 | rowspan="9" | TSMC 28HP | {{unk}} | {{unk}} | 512:24:8 (4) | PCIe 3.0 x8 | ? | 1072 | 1176 | 1800 | 1 / 2 | 14.4 | DDR3 / GDDR5 | 64 | 8.5 9.4 | 25.7 28.2 | 1,097.7 1,204.2 | 34.3 37.6 | 33 | rowspan="1" {{No}} | {{okay|OEM}} |- ! style="text-align:left;" | GeForce GTX 950<ref>{{cite web |url=http://www.geforce.com/hardware/desktop-gpus/geforce-gtx-950/specifications |title=GTX 950 {{pipe}} Specifications |publisher=GeForce |access-date=2015-12-11 |archive-url=https://web.archive.org/web/20151212232816/http://www.geforce.com/hardware/desktop-gpus/geforce-gtx-950/specifications |archive-date=2015-12-12 |url-status=live }}</ref> | August 20, 2015 | GM206-250 | rowspan="3" | 2.94 | rowspan="3" | 227 | 768:48:32 (6) | rowspan="8" | PCIe 3.0 x16 | rowspan="4" | 1 | 1024 | 1188 | 6600 | rowspan="2" | 2 | 105.7 | rowspan="8" | GDDR5 | rowspan="3" | 128 | 32.7 38.0 | 49.1 57.0 | 1,572.8 1,824.7 | 49.1 57.0 | 90 (75{{efn|name=GTX950NPC|Some GTX950 cards were released without power connector powered only by PCIe slot. These had limited power consumption and TDP to 75W.<ref>{{Cite web|url=https://www.anandtech.com/show/10250/gigabyte-adds-geforce-gtx-950-with-75w-power-consumption-to-lineup|archive-url=https://web.archive.org/web/20160415034354/http://www.anandtech.com/show/10250/gigabyte-adds-geforce-gtx-950-with-75w-power-consumption-to-lineup|url-status=dead|archive-date=April 15, 2016|title=GIGABYTE Adds 75W GeForce GTX 950 to Lineup|first=Anton|last=Shilov|website=www.anandtech.com|accessdate=May 16, 2024}}</ref> }}) | rowspan="4" | 2-way SLI | 159 |- ! style="text-align:left;" | GeForce GTX 950 (OEM)<ref>{{cite web|url=https://www.geforce.com/hardware/desktop-gpus/geforce-gtx-950-oem/specifications|title=GeForce GTX 950 (OEM) {{pipe}} Specifications {{pipe}} GeForce|website=geforce.com|access-date=2019-01-09|archive-url=https://web.archive.org/web/20180923200607/https://www.geforce.com/hardware/desktop-gpus/geforce-gtx-950-oem/specifications|archive-date=2018-09-23|url-status=live}}</ref> | {{unk}} | GM206 | rowspan="2" | 1024:64:32 (8) | 935 | {{unk}} | 5000 | 80.0 | 29.9   | 59.8   | 1,914.9 , | 59.8   | {{unk}} | {{okay|OEM}} |- ! style="text-align:left;" | GeForce GTX 960<ref>{{cite web |url=http://www.geforce.com/hardware/desktop-gpus/geforce-gtx-960/specifications |title=GTX 960 {{pipe}} Specifications |publisher=GeForce |access-date=2015-12-11 |archive-url=https://web.archive.org/web/20151212024030/http://www.geforce.com/hardware/desktop-gpus/geforce-gtx-960/specifications |archive-date=2015-12-12 |url-status=live }}</ref> | January 22, 2015 | GM206-300 | 1127 | 1178 | 7000 | 2 4{{efn|name=GTX9604GB| Some manufacturers produced 4 GB versions of GTX 960. These were often criticized as useless move, as titles that would use so much VRAM and actually gain advantage over 2 GB versions, would already run too slow on those resolutions and settings, as GTX960 didn't have enough compute power and memory bandwidth to handle it.<ref>{{cite web |title=Nvidia GeForce GTX 960 2GB vs 4GB review |website=Eurogamer |date=18 October 2015 |url=https://www.eurogamer.net/articles/digitalfoundry-2015-nvidia-geforce-gtx-960-2gb-vs-4gb-review }}</ref> }} | 112.1 | 36.0 37.6 | 72.1 75.3 | 2,308.0 2,412.5 | 72.1 75.3 | 120 | 199 |- ! style="text-align:left;" | GeForce GTX 960 (OEM)<ref>{{cite web |url=https://www.geforce.com/hardware/desktop-gpus/geforce-gtx-960-oem/specifications |title=GeForce GTX 960 (OEM) {{pipe}} Specifications {{pipe}} GeForce |website=geforce.com |access-date=2019-01-09 |archive-url=https://web.archive.org/web/20180923200618/https://www.geforce.com/hardware/desktop-gpus/geforce-gtx-960-oem/specifications |archive-date=2018-09-23 |url-status=live }}</ref> | {{unk}} | GM204 | rowspan="3" | 5.2 | rowspan="3" | 398 | 1280:80:48 (10) | 924 | {{unk}} | 5000 | 3 | 120.0 | 192 | 44.3   | 73.9   | 2,365.4 , | 73.9   | {{unk}} | {{okay|OEM}} |- ! style="text-align:left;" | GeForce GTX 970<ref>{{cite web |url=http://www.geforce.com/hardware/desktop-gpus/geforce-gtx-970/specifications |title=GTX 970 {{pipe}} Specifications |publisher=GeForce |access-date=2015-12-11 |archive-url=https://web.archive.org/web/20151207185709/http://www.geforce.com/hardware/desktop-gpus/geforce-gtx-970/specifications |archive-date=2015-12-07 |url-status=live }}</ref> | September 18, 2014 | GM204-200 | 1664:104:56 (13) | 1.75 | 1050 | 1178 | rowspan="4" | 7000 | 3.5 + 0.5{{efn|name=GTX960MemoryMess|For accessing its memory, the GTX 970 stripes data across 7 of its 8 32-bit physical memory lanes, at 196 GB/s. The last 1/8 of its memory (0.5 GB on a 4 GB card) is accessed on a non-interleaved solitary 32-bit connection at 28 GB/s, one seventh the speed of the rest of the memory space. Because this smaller memory pool uses the same connection as the 7th lane to the larger main pool, it contends with accesses to the larger block reducing the effective memory bandwidth not adding to it as an independent connection could.<ref>{{cite news |last1=Wasson |first1=Scott |date=January 26, 2015 |title=Nvidia: the GeForce GTX 970 works exactly as intended, A look inside the card's unusual memory config |url=http://techreport.com/review/27724/nvidia-the-geforce-gtx-970-works-exactly-as-intended |newspaper=The Tech Report |page=1 |access-date=2015-01-26 |archive-url=https://web.archive.org/web/20150128051624/http://techreport.com/review/27724/nvidia-the-geforce-gtx-970-works-exactly-as-intended |archive-date=January 28, 2015 |url-status=live }}</ref>}} | 196.3 + 28.0{{efn|name=GTX960MemoryMess}} | 224 + 32{{efn|name=GTX960MemoryMess}} | 58.8 65.9 | 109.2 122.5 | 3,494.4 3,920.3 | 109.2 122.5 | 145 | rowspan="4" | 4-way SLI | 329 |- ! style="text-align:left;" | GeForce GTX 980<ref>{{cite web |url=http://www.geforce.com/hardware/desktop-gpus/geforce-gtx-980/specifications |title=GTX 980 {{pipe}} Specifications |publisher=GeForce |access-date=2015-12-11 |archive-url=https://web.archive.org/web/20151208184430/http://www.geforce.com/hardware/desktop-gpus/geforce-gtx-980/specifications |archive-date=2015-12-08 |url-status=live }}</ref> | September 18, 2014 | GM204-400 | 2048:128:64 (16) | 2 | 1126 | 1216 | 4 | 224.3 | 256 | 72.0 77.8 | 144.1 155.6 | 4,612.0 4,980.7 | 144.1 155.6 | 165 | 549 |- ! style="text-align:left;" | GeForce GTX 980 Ti<ref>{{cite web |url=http://www.geforce.com/hardware/desktop-gpus/geforce-gtx-980-ti/specifications |title=GTX 980 Ti {{pipe}} Specifications |publisher=GeForce |access-date=2015-12-11 |archive-url=https://web.archive.org/web/20151211174512/http://www.geforce.com/hardware/desktop-gpus/geforce-gtx-980-ti/specifications |archive-date=2015-12-11 |url-status=live }}</ref> | June 1, 2015 | GM200-310 | rowspan="2" | 8 | rowspan="2" | 601 | 2816:176:96 (22) | rowspan="2" | 3 | rowspan="2" | 1000 | rowspan="2" | 1075 | 6 | rowspan="2" | 336.5 | rowspan="2" | 384 | rowspan="2" | 96.0 103.2 | 176.0 189.2 | 5,632.0 6,054.4 | 176.0 189.2 | rowspan="2" | 250 | 649 |- ! style="text-align:left;" | GeForce GTX Titan X<ref>{{cite web |url=http://www.geforce.com/hardware/desktop-gpus/geforce-gtx-titan-x/specifications |title=GTX TITAN X {{pipe}} Specifications |publisher=GeForce |access-date=2015-12-11 |archive-url=https://web.archive.org/web/20151205173930/http://www.geforce.com/hardware/desktop-gpus/geforce-gtx-titan-x/specifications |archive-date=2015-12-05 |url-status=live }}</ref> | March 17, 2015 | GM200-400 | 3072:192:96 (24) | 12 | 192.0 206.4 | 6,144.0 6,604.8 | 192.0 206.4 | 999 |} {{notelist|refs= {{efn|name=CoreConfig|Main shader processors: texture mapping units: render output units (streaming multiprocessors)}} {{efn|name=PixelFillrate|Pixel fillrate is calculated as the number of ROPs multiplied by the respective core clock speed.}} {{efn|name=TextureFillrate|Texture fillrate is calculated as the number of TMUs multiplied by the respective core clock speed.}} {{efn|name=ProcessingPower|To calculate the processing power see Maxwell (microarchitecture)#Performance.}} {{efn|name=PerfValues|Base clock, Boost clock}} }}

===GeForce 900M (9xxM) series=== Some implementations may use different specifications.

{{row hover highlight}} {{sort under}} {| class="wikitable sortable sort-under mw-datatable" style="font-size: 85%; text-align: center;" |- ! rowspan="2" | Model ! rowspan="2" | Launch ! rowspan="2" | Code name ! rowspan="2" | Fab (nm) ! rowspan="2" | Transistors (million) ! rowspan="2" | Die size (mm2) ! rowspan="2" | Bus interface ! rowspan="2" | Core config{{efn|name=a}} ! colspan="3" | Clock speeds ! colspan="2" | Fillrate ! colspan="4" | Memory ! colspan="4" | API support (version) ! colspan="2" | Processing power (GFLOPS) ! rowspan="2" | TDP (watts) ! rowspan="2" | SLI support{{efn|name="SLI"}} |- ! Base core clock (MHz) ! Boost core clock (MHz) ! Memory (MT/s) ! Pixel (GP/s){{efn|name="pixel fillrate"|Pixel fillrate is calculated as the lowest of three numbers: number of ROPs multiplied by the base core clock speed, number of rasterizers multiplied by the number of fragments they can generate per rasterizer multiplied by the base core clock speed, and the number of streaming multiprocessors multiplied by the number of fragments per clock that they can output multiplied by the base clock rate.<ref name="techreport.com" />}} ! Texture (GT/s){{efn|name="texture fillrate"}} ! Size (MiB) ! Bandwidth (GB/s) ! Type ! Bus width (bit) ! DirectX ! OpenGL ! OpenCL ! Vulkan ! Single precision{{efn|name=single FLOPS"}} ! Double precision{{efn|name="double FLOPS"|Double precision performance of the '''Maxwell''' chips' are 1/32 of single-precision performance.<ref name="AnandTech GTX 980">{{cite news |last1=Smith |first1=Ryan |date=September 18, 2014 |title=The NVIDIA GeForce GTX 980 Review: Maxwell Mark 2 |url=http://www.anandtech.com/show/8526/nvidia-geforce-gtx-980-review |archive-url=https://web.archive.org/web/20140922001256/http://www.anandtech.com/show/8526/nvidia-geforce-gtx-980-review |url-status=dead |archive-date=September 22, 2014 |newspaper=AnandTech |page=1 |access-date=September 19, 2014}}</ref><ref name="Anandtech Titan X">{{cite web|url=http://anandtech.com/show/9059/the-nvidia-geforce-gtx-titan-x-review|archive-url=https://archive.today/20150317191144/http://anandtech.com/show/9059/the-nvidia-geforce-gtx-titan-x-review|url-status=dead|archive-date=March 17, 2015|title=The NVIDIA GeForce GTX Titan X Review|author=Ryan Smith|work=anandtech.com}}</ref>}} |- ! style="text-align:left;" rowspan="2" | GeForce 910M<ref>{{cite web|url=https://www.geforce.com/hardware/notebook-gpus/geforce-910m/specifications|title=GeForce 910M - Specifications - GeForce|work=geforce.com}}</ref><ref>{{cite web |url=https://www.techpowerup.com/gpu-specs/geforce-910m.c2764 |title=GeForce 910M |access-date=October 27, 2016 |archive-url=https://web.archive.org/web/20170624062933/https://www.techpowerup.com/gpu-specs/geforce-910m.c2764 |archive-date=June 24, 2017 |url-status=dead }}</ref><ref>{{cite web |url=https://www.techpowerup.com/gpu-specs/geforce-910m.c2764 |title=NVIDIA GeForce 910M Specs | TechPowerUp GPU Database |publisher=Techpowerup.com |date=August 22, 2022 |accessdate=2026-02-19}}</ref> | Aug 18, 2015 | GF117{{efn|name="nohwenc"|Lacks hardware video encoder}} | rowspan="19" | 28 | 585 | 116 | rowspan="6" | PCIe 3.0 x8 | 96:16:8 | 775 | 1550 | rowspan="6" | 1800 | 3.1 | 12.4 | 1024 | rowspan="6" | 14.4 | rowspan="4" | GDDR3 | rowspan="7" | 64 | rowspan="15" |12.0 (11_0)<ref name="D3D11.3"/><ref name="11.3And12.0RenderingFeaturesAreEquivalent"/> | rowspan="19" | 4.6 | 1.1 | {{n/a}} | 297.6 | 1/12 of SP | rowspan="4" | 33 | rowspan="6" | No |- | March 15, 2015 | GK208 | {{unk}} | 87 | 384:16:8 | 575 | 575 | 5.13 | 9.2 | 2048 | 1.2 | 1.1 | 441.6 | 18.4 |- ! style="text-align:left;" rowspan="2" | GeForce 920M<ref>{{cite web|url=http://www.geforce.com/hardware/notebook-gpus/geforce-920m/specifications|title=GeForce 920M - Specifications - GeForce|work=geforce.com}}</ref><ref>{{cite web|url=http://www.techpowerup.com/gpudb/2675/geforce-920m.html|title=NVIDIA GeForce 920M|work=TechPowerUp|access-date=March 17, 2015|archive-url=https://web.archive.org/web/20160304061000/http://www.techpowerup.com/gpudb/2675/geforce-920m.html|archive-date=March 4, 2016|url-status=dead}}</ref><ref>{{cite web|url=https://www.techpowerup.com/gpu-specs/geforce-920m.c2646|title=NVIDIA GeForce 920M|work=TechPowerUp |accessdate=2026-02-19}}</ref> | rowspan="2" | March 13, 2015 | GF117{{efn|name="nohwenc"}} | 585 | 116 | 96:16:8 | 775 | 1550 | 3.1 | 12.4 | 1024 | 1.1 | {{n/a}} | 297.6 | 1/12 of SP |- | GK208 | {{unk}} | 87 | 384:32:16 | 954 | 954 | 7.6 | 30.5 | 2048 | rowspan="3" | 1.2 | rowspan="3" | 1.1 | 732.7 | 22.9 |- ! style="text-align:left;" | GeForce 920MX<ref>{{cite web|url=https://www.geforce.com/hardware/notebook-gpus/geforce-920mx/specifications|title=GeForce 920MX - Specifications - GeForce|work=geforce.com}}</ref><ref>{{cite web |url=https://www.techpowerup.com/gpudb/2826/geforce-920mx |title=NVIDIA GeForce 920MX Specs | TechPowerUp GPU Database |publisher=Techpowerup.com |date=August 22, 2022 |accessdate=2022-08-22}}</ref> | March 2016 | rowspan="3" | GM108{{efn|name="nohwcodec"|Lacks hardware video encoder and decoder}} | rowspan="2" | 1870 | rowspan="2" | 148 | 256:24:8 | 1072 | 1176 | 8.58 | 25.7 | 2048 | DDR3 GDDR5 | 549 | 1/32 of SP | 16 |- ! style="text-align:left;" | GeForce 930M<ref>{{cite web|url=https://www.geforce.com/hardware/notebook-gpus/geforce-930m/specifications|title=GeForce 930M - Specifications - GeForce|work=geforce.com}}</ref><ref>{{cite web|url=https://www.techpowerup.com/gpudb/2644/geforce-930m.html|title=NVIDIA GeForce 930M|work=TechPowerUp}}</ref> | March 13, 2015 | 384:24:8 | 928 | 941 | 7.4 | 22.3 | 2048 | DDR3 | 712.7 | 22.3 | 33 |- ! style="text-align:left;" | GeForce 930MX<ref>{{cite web|url=http://www.geforce.com/hardware/notebook-gpus/geforce-930mx/specifications|title=GeForce 930MX - Specifications - GeForce|work=geforce.com}}</ref><ref>{{cite web |url=https://www.techpowerup.com/gpudb/2825/geforce-930mx |title=NVIDIA GeForce 930MX Specs | TechPowerUp GPU Database |publisher=Techpowerup.com |date=August 22, 2022 |accessdate=2022-08-22}}</ref> | March 1, 2016 | {{unk}} | {{unk}} | PCIe 3.0 x8 | 384:24:8 | 952 | 1020 | 2000 | {{unk}} | {{unk}} | 2048 | {{unk}} | DDR3 GDDR5 | {{unk}} | {{unk}} | {{unk}} | {{unk}} | {{unk}} | {{unk}} |- ! style="text-align:left;" rowspan="2" | GeForce 940M<ref>{{cite web|url=http://www.geforce.com/hardware/notebook-gpus/geforce-940m/specifications|title=GeForce 940M - Specifications - GeForce|work=geforce.com}}</ref><ref>{{cite web|url=http://www.techpowerup.com/gpudb/2648/geforce-940m.html|title=NVIDIA GeForce 940M|work=TechPowerUp}}</ref><ref>{{cite web|url=http://www.techpowerup.com/gpudb/2643/geforce-940m.html|title=NVIDIA GeForce 940M|work=TechPowerUp}}</ref> | rowspan="2" | March 13, 2015 | GM107 | 1870 | 148 | PCIe 3.0 x16 | 640:40:16 | rowspan="2" | 1029 | rowspan="2" | 1100 | rowspan="3" | 2002 | 16.5 | 41.2 | rowspan="2" | 2048 | rowspan="2" | 16 - 80.2 | rowspan="3" | GDDR5 DDR3 | 128 | rowspan="6" | 1.2 | rowspan="5" | 1.1 | 1317 | 41.1 | 75 | rowspan="3" | No |- | rowspan="2" | GM108{{efn|name="nohwcodec"}} | {{unk}} | {{unk}} | rowspan="2" | PCIe 3.0 x8 | 384:24:8 | 8.2 | 24.7 | rowspan="2" | 64 | 790.3 | 24.7 | 33 |- ! style="text-align:left;" | GeForce 940MX<ref>{{cite web|url=http://www.geforce.com/hardware/notebook-gpus/geforce-940mx/specifications|title=GeForce 940MX - Specifications - GeForce|work=geforce.com}}</ref><ref>{{cite web| url = https://www.techpowerup.com/gpudb/2797/geforce-940mx | title = NVIDIA GeForce 940MX | work = TechPowerUp GPU Database | access-date = December 16, 2017}}</ref> | March 10, 2016 | rowspan="2" | 1870 | rowspan="2" | 148 | 384:24:8 | 1122 | 1242 | 8.98 | 26.93 | 2048 4096 | 16.02 (DDR3) 40.1 (GDDR5) | 861.7 | {{unk}} | 23 |- ! style="text-align:left;" rowspan="2" | GeForce 945M<ref>{{cite web|url=http://www.geforce.com/hardware/notebook-gpus/geforce-945m/specifications|title=GeForce 945M - Specifications - GeForce|work=geforce.com}}</ref><ref>{{cite web |url=https://www.techpowerup.com/gpudb/2773/geforce-945m |title=NVIDIA GeForce 945M Specs | TechPowerUp GPU Database |publisher=Techpowerup.com |date=August 22, 2022 |accessdate=2022-08-22}}</ref><ref>{{cite web |url=https://www.techpowerup.com/gpudb/2836/geforce-945m |title=NVIDIA GeForce 945M Specs | TechPowerUp GPU Database |publisher=Techpowerup.com |date=August 22, 2022 |accessdate=2022-08-22}}</ref> | rowspan="2" | 2015 | GM107 | {{dunno}} | 640:40:16 | 1029 | 1085 | rowspan="2" {{dunno}} | 16.46 | 41.2 | rowspan="2" {{dunno}} | rowspan="2" {{dunno}} | rowspan="2" | DDR3 GDDR5 | 128 | 1,317.1 | rowspan="2" {{dunno}} | 75 | rowspan="2" {{dunno}} |- | rowspan="2" | GM108{{efn|name="nohwcodec"}} | {{dunno}} | {{dunno}} | rowspan="2" | PCIe 3.0 x8 | 384:24:8 | 1122 | 1242 | 8.98 | 26.93 | rowspan="2" | 64 | 861.7 | 23 |- ! style="text-align:left;" | GeForce GT 945A<ref>NVIDIA™ GeForceGT 945A (1GB GDDR5) user-selectable by application via NVIDIA Control Panel http://store.hp.com/us/en/ContentView?catalogId=10051&langId=-1&storeId=10151&eSpotName=Sprout-Pro#!</ref><ref>{{cite web |url=https://www.techpowerup.com/gpudb/2813/geforce-945a |title=NVIDIA GeForce 945A Specs | TechPowerUp GPU Database |publisher=Techpowerup.com |date=August 22, 2022 |accessdate=2022-08-22}}</ref> | March 13, 2015 | {{unk}} | {{unk}} | 384:24:8 | 1072 | 1176 | 1800 | 8.58 | 25.73 | 2048 | 14.4 | DDR3 | {{unk}} | {{unk}} | {{unk}} | 33 | {{unk}} |- ! style="text-align:left;" | GeForce GTX 950M<ref>{{cite web|url=http://www.geforce.com/hardware/notebook-gpus/geforce-gtx-950m/specifications|title=GeForce GTX 950M - Specifications - GeForce|work=geforce.com}}</ref><ref>{{cite web|url=http://www.techpowerup.com/gpudb/2642/geforce-gtx-950m.html|title=NVIDIA GeForce GTX 950M|work=TechPowerUp}}</ref> | rowspan="2" | March 13, 2015 | rowspan="2" | GM107 | rowspan="2" | 1870 | rowspan="2" | 148 | rowspan="6" | PCIe 3.0 x16 | 640:40:16 | 914 | 1085 | rowspan="2" | 5012 | 14.6 | 36.6 | 2048(GDDR5) 4096(DDR3) | 80(GDDR5) 32(DDR3) | DDR3 GDDR5 | rowspan="3" | 128 | rowspan="6" | 1.2<ref name="TPUDB980">{{cite web|url=http://www.techpowerup.com/gpudb/2621/geforce-gtx-980.html|title=NVIDIA GeForce GTX 980|work=TechPowerUp}}</ref> | rowspan="6" | 1.1 | 1170 | 36.56 | 75 | rowspan="2" | No |- ! style="text-align:left;" | GeForce GTX 960M<ref>{{cite web|url=http://www.geforce.com/hardware/notebook-gpus/geforce-gtx-960m/specifications|title=GeForce GTX 960M - Specifications - GeForce|work=geforce.com}}</ref><ref>{{cite web|url=http://www.techpowerup.com/gpudb/2635/geforce-gtx-960m.html|title=NVIDIA GeForce GTX 960M|work=TechPowerUp}}</ref> | 640:40:16 | 1029 | 1085 | 16.5 | 41.2 | rowspan="2" | 2048 4096 | rowspan="2" | 80 | rowspan="5" |GDDR5 | 1317 | 41.16 | 65 |- ! style="text-align:left;" | GeForce GTX 965M<ref>{{cite web|url=http://www.geforce.com/hardware/notebook-gpus/geforce-gtx-965m/specifications|title=GeForce GTX 965M - Specifications - GeForce|work=geforce.com}}</ref><ref>{{cite web|url=http://www.techpowerup.com/gpudb/2634/geforce-gtx-965m.html|title=NVIDIA GeForce GTX 965M|work=TechPowerUp}}</ref> | January 5, 2015 | rowspan="4" | GM204 | rowspan="4" | 5200 | rowspan="4" | 398 | 1024:64:32 | 924 | 950 | 5000 | 30.2 | 60.4 | rowspan="4" | 12.0 (12_1)<ref name="D3D11.3"/><ref name="11.3And12.0RenderingFeaturesAreEquivalent"/> | 1945 | 60.78 | 60 | rowspan="4" | Yes |- ! style="text-align:left;" | GeForce GTX 970M<ref>{{cite web|url=http://www.geforce.com/hardware/notebook-gpus/geforce-gtx-970m/specifications|title=GeForce GTX 970M - Specifications - GeForce|work=geforce.com}}</ref> | rowspan="2" | October 7, 2014 | 1280:80:48 | 924 | 993 | rowspan="2" | 5012 | 37.0 | 73.9 | 3072 6144 | 120 | 192 | 2365 | 73.9 | 75 |- ! style="text-align:left;" | GeForce GTX 980M<ref>{{cite web|url=http://www.geforce.com/hardware/notebook-gpus/geforce-gtx-980m/specifications|title=GeForce GTX 980M - Specifications - GeForce|work=geforce.com}}</ref> | 1536:96:64 | 1038 | 1127 | 49.8 | 99.6 | rowspan="2" | 4096 8192 | 160 | 256<ref name="hardware.fr">{{cite web|url=https://www.hardware.fr/focus/106/gtx-970-3-5-go-224-bit-lieu-4-go-256-bit.html|title=GTX 970: 3.5 Go et 224-bit au lieu de 4 Go et 256-bit ?|first=Damien |last=Triolet |language=fr |work=Hardware.FR |date=February 4, 2016 |access-date=May 27, 2016}}</ref> | 3189 | 99.6 | 100 |} {{notelist|refs= {{efn|name=a|Shader Processors: Texture mapping units: Render output units}} {{efn|name="texture fillrate"|Texture fillrate is calculated as the number of TMUs multiplied by the base core clock speed.}} {{efn|name=single FLOPS"|Single precision performance is calculated as 2 times the number of shaders multiplied by the base core clock speed.}} {{efn|name="SLI"|A maximum of 2 dual-GPU cards can be connected in tandem for a 4-way SLI configuration as dual-GPU cards feature on-board 2-way SLI.}} }}

==Advertising controversy== ===GTX 970 hardware specifications=== thumb|Nvidia GeForce GTX 970 PCB and die Issues with the GeForce GTX 970's specifications were first brought up by users when they found out that the cards, while featuring 4 GB of memory, rarely accessed memory over the 3.5 GB boundary. Further testing and investigation eventually led to Nvidia issuing a statement that the card's initially announced specifications had been altered without notice before the card was made commercially available, and that the card took a performance hit once memory over the 3.5 GB limit were put into use.<ref>{{cite web|url=http://www.pcper.com/reviews/Graphics-Cards/NVIDIA-Discloses-Full-Memory-Structure-and-Limitations-GTX-970|title=NVIDIA Discloses Full Memory Structure and Limitations of GTX 970|publisher=PCPer|access-date=January 28, 2015|archive-url=https://web.archive.org/web/20150225180420/http://www.pcper.com/reviews/Graphics-Cards/NVIDIA-Discloses-Full-Memory-Structure-and-Limitations-GTX-970|archive-date=February 25, 2015|url-status=dead}}</ref><ref>{{cite web|url=http://wccftech.com/nvidia-geforce-gtx-970-memory-issue-fully-explained/|title=GeForce GTX 970 Memory Issue Fully Explained – Nvidia's Response|date=January 24, 2015|publisher=WCFTech}}</ref><ref>{{cite web|url=https://www.pcgamer.com/why-nvidias-gtx-970-slows-down-using-more-than-35gb-vram/|title=Why Nvidia's GTX 970 slows down when using more than 3.5GB VRAM|date=January 26, 2015|publisher=PCGamer}}</ref>

The card's back-end hardware specifications, initially announced as being identical to those of the GeForce GTX 980, differed in the amount of L2 cache (1.75 MB versus 2 MB in the GeForce GTX 980) and the number of ROPs (56 versus 64 in the 980). Additionally, it was revealed that the card was designed to access its memory as a 3.5 GB section, plus a 0.5 GB one, access to the latter being 7 times slower than the first one.<ref>{{cite web|url=http://www.anandtech.com/show/8935/geforce-gtx-970-correcting-the-specs-exploring-memory-allocation|archive-url=https://web.archive.org/web/20150128095356/http://www.anandtech.com/show/8935/geforce-gtx-970-correcting-the-specs-exploring-memory-allocation|url-status=dead|archive-date=January 28, 2015|title=GeForce GTX 970: Correcting The Specs & Exploring Memory Allocation|publisher=AnandTech}}</ref> The company then went on to promise a specific driver modification in order to alleviate the performance issues produced by the cutbacks suffered by the card.<ref>{{cite web|url=http://wccftech.com/nvidia-working-driver-geforce-gtx-970-tune-memory-allocation-problems-improve-performance/|title=NVIDIA Working on New Driver For GeForce GTX 970 To Tune Memory Allocation Problems and Improve Performance|date=January 28, 2015|publisher=WCFTech}}</ref> However, Nvidia later clarified that the promise had been a miscommunication and there would be no specific driver update for the GTX 970.<ref>{{cite web|url=https://www.pcworld.com/article/2876802/nvidia-plans-geforce-gtx-970-driver-update-for-memory-performance-concerns.html/|title= NVIDIA clarifies no driver update for GTX 970 specifically|date= January 29, 2015|publisher=PC World}}</ref> Nvidia claimed that it would assist customers who wanted refunds in obtaining them.<ref>{{cite web|url=https://www.pcper.com/news/Graphics-Cards/NVIDIA-Plans-Driver-Update-GTX-970-Memory-Issue-Help-Returns|title=NVIDIA Plans Driver Update for GTX 970 Memory Issue, Help with Returns|work=pcper.com|date=January 28, 2015}}</ref> On February 26, 2015, Nvidia CEO Jen-Hsun Huang went on record in Nvidia's official blog to apologize for the incident.<ref>{{cite web|url=https://www.pcgamer.com/nvidia-ceo-addresses-gtx-970-controversy/|title=Nvidia CEO addresses GTX 970 controversy|publisher=PCGamer|date=February 26, 2015}}</ref> In February 2015 a class-action lawsuit alleging false advertising was filed against Nvidia and Gigabyte Technology in the U.S. District Court for Northern California.<ref name="970lawsuit">{{cite news|url=https://www.pcgamer.com/nvidia-faces-false-advertising-lawsuit-over-gtx-970-specs/|title=Nvidia faces false advertising lawsuit over GTX 970 specs|last=Chalk|first=Andy|date=February 22, 2015|work=PC Gamer|access-date=March 27, 2015}}</ref><ref>{{cite news|url=https://www.pcworld.com/article/2887234/nvidia-hit-with-false-advertising-suit-over-gtx-970-performance.html|title=Nvidia hit with false advertising suit over GTX 970 performance|last=Niccolai|first=James|date=February 20, 2015|work=PC World|access-date=March 27, 2015}}</ref>

Nvidia revealed that it is able to disable individual units, each containing 256KB of L2 cache and 8 ROPs, without disabling whole memory controllers.<ref name="AnandTechCorrectionPage2">{{cite web|url=http://www.anandtech.com/show/8935/geforce-gtx-970-correcting-the-specs-exploring-memory-allocation/2|archive-url=https://web.archive.org/web/20150128132040/http://www.anandtech.com/show/8935/geforce-gtx-970-correcting-the-specs-exploring-memory-allocation/2|url-status=dead|archive-date=January 28, 2015|title=Diving Deeper: The Maxwell 2 Memory Crossbar & ROP Partitions - GeForce GTX 970: Correcting The Specs & Exploring Memory Allocation|author=Ryan Smith|work=anandtech.com}}</ref> This comes at the cost of dividing the memory bus into high speed and low speed segments that cannot be accessed at the same time unless one segment is reading while the other segment is writing because the L2/ROP unit managing both of the GDDR5 controllers shares the read return channel and the write data bus between the two GDDR5 controllers and itself.<ref name="AnandTechCorrectionPage2"/> This is used in the GeForce GTX 970, which therefore can be described as having 3.5 GB in its high speed segment on a 224-bit bus and 0.5 GB in a low speed segment on a 32-bit bus.<ref name="AnandTechCorrectionPage2"/>

On July 27, 2016, Nvidia agreed to a preliminary settlement of the U.S. class action lawsuit,<ref name="970lawsuit"/> offering a $30 refund on GTX 970 purchases. The agreed upon refund represents the portion of the cost of the storage and performance capabilities the consumers assumed they were obtaining when they purchased the card.<ref>{{cite news|url=https://topclassactions.com/lawsuit-settlements/lawsuit-news/340705-nvidia-settles-graphics-card-false-advertising-class-action|title=Nvidia settles class action lawsuit|date=July 27, 2016|work=Top Class Actions|access-date=July 27, 2016}}</ref>

== Support == Driver 368.81 is the last driver to support Windows XP/Windows XP 64-bit.{{fact|date=September 2024}}

32-bit drivers for 32-bit operating systems were discontinued after the release of driver 391.35 in March 2018.<ref>{{cite web|url=http://nvidia.custhelp.com/app/answers/detail/a_id/4604/|title = Support Plan for 32-bit and 64-bit Operating Systems | NVIDIA}}</ref>

Notebook GPUs based on the Kepler architecture moved to legacy support in April 2019 and stopped receiving critical security updates after April 2020.<ref>{{cite news |url=https://www.techspot.com/news/79115-nvidia-end-support-mobile-kepler-gpus-starting-april.html |publisher=Techspot |author=Eric Hamilton |date=March 9, 2019 |title=Nvidia to end support for mobile Kepler GPUs starting April 2019}}</ref><ref>{{cite web |url=https://nvidia.custhelp.com/app/answers/detail/a_id/4780 |publisher=Nvidia |title=List of Kepler series GeForce Notebook GPUs}}</ref> The Nvidia GeForce 910M and 920M from the 9xxM GPU family are affected by this change.

Nvidia announced that after the release of the 470 driver branch, it would transition driver support for the Windows 7 and Windows 8.1 operating systems to legacy status and continue to provide critical security updates for these operating systems through September 2024.<ref>{{cite web|url=https://nvidia.custhelp.com/app/answers/detail/a_id/5201|title = Support Plan for Windows 7 and Windows 8/8.1 | NVIDIA}}</ref>

In December 2025, Nvidia discontinued Game Ready Driver support for the Maxwell, Pascal, and Volta architectures, which includes the GeForce 900 series. Security updates are scheduled to be continued until October 2028, as per Nvidia.<ref>{{Cite web |title=NVIDIA officially ends GeForce GTX 10 & GTX 900 Game Ready driver support, brings back 32-bit PhysX support for RTX 50 series |url=https://videocardz.com/newz/nvidia-officially-ends-geforce-gtx-10-gtx-900-game-ready-driver-support-brings-back-32-bit-physx-support-for-rtx-50-series |access-date=2026-02-19 |website=VideoCardz.com |language=en-US}}</ref><ref>{{Cite web |title=NVIDIA Support |url=https://nvidia.custhelp.com/app/answers/detail/a_id/5676/~/support-plan-for-maxwell%2C-pascal%2C-and-volta-series-geforce-gpus. |access-date=2026-02-19 |website=nvidia.custhelp.com |language=en-US}}</ref> On July 1, 2025, Nvidia announced that driver branch 580 will be the last to support these architectures.<ref>{{Cite web |title=NVIDIA confirms end of support for GeForce GTX 700, 900 and 10 series after 580 drivers |url=https://videocardz.com/newz/nvidias-next-major-gpu-driver-branch-to-drop-support-for-geforce-gtx-700-900-and-10-series |access-date=2025-07-23 |website=VideoCardz.com |language=en-US}}</ref>

==Gallery== <gallery class="center" widths="220"> File:Gigabyte GTX 970 Windforce.jpg|Nvidia GeForce GTX 970 'Windforce' edition. This particular model manufactured by Nvidia board-partner, Gigabyte. File:EVGA GTX 960 2GB 1 2019-03-21.jpg|Nvidia GeForce GTX 960 'ACX 2.0' edition. This particular model manufactured by Nvidia board-partner, EVGA. File:CyberPower Syber Steam Machine I (22984998030).png|Nvidia GeForce GTX 950. This particular model manufactured by Nvidia board-partner, MSI. </gallery>

==References== {{Reflist|30em}}

==External links== * [https://web.archive.org/web/20170721113746/http://international.download.nvidia.com/geforce-com/international/pdfs/GeForce_GTX_980_Whitepaper_FINAL.PDF GeForce GTX 980 Whitepaper] * [https://www.geforce.com/whats-new/articles/nvidia-geforce-gtx-titan-x The Ultimate GPU, TITAN X. Available Now] * [http://www.geforce.com/whats-new/articles/introducing-the-geforce-gtx-980-ti Introducing The GeForce GTX 980 Ti. Play The Future] * [https://www.geforce.com/whats-new/articles/maxwell-architecture-gtx-980-970 Introducing The Amazing New GeForce GTX 980 & 970] * [http://www.geforce.com/whats-new/articles/introducing-the-geforce-gtx-960 Introducing The $199 GeForce GTX 960: High-End Performance & High-End Features On A Mid-Range GPU] * [https://www.geforce.com/whats-new/articles/introducing-the-geforce-gtx-950 Gain The Competitive Edge With The All-New GeForce GTX 950] * [https://www.geforce.com/whats-new/articles/geforce-gtx-900m-laptops-available-now GeForce GTX 900M: The World’s Most Advanced Gaming Notebooks] * [http://www.geforce.com/hardware/desktop-gpus/geforce-gtx-titan-x GeForce GTX TITAN X] * [http://www.geforce.com/hardware/desktop-gpus/geforce-gtx-980-ti GeForce GTX 980 Ti] * [http://www.geforce.com/hardware/desktop-gpus/geforce-gtx-980 GeForce GTX 980] * [http://www.geforce.com/hardware/desktop-gpus/geforce-gtx-970 GeForce GTX 970] * [http://www.geforce.com/hardware/desktop-gpus/geforce-gtx-960 GeForce GTX 960] * [http://www.geforce.com/hardware/desktop-gpus/geforce-gtx-950 GeForce GTX 950] * [http://www.geforce.com/hardware/notebook-gpus/geforce-gtx-980m GeForce GTX 980M] * [http://www.geforce.com/hardware/notebook-gpus/geforce-gtx-970m GeForce GTX 970M] * [http://www.geforce.com/hardware/notebook-gpus/geforce-gtx-965m GeForce GTX 965M] * [http://www.geforce.com/hardware/notebook-gpus/geforce-gtx-960m GeForce GTX 960M] * [http://www.geforce.com/hardware/notebook-gpus/geforce-gtx-950m GeForce GTX 950M] * [https://www.geforce.com/hardware/notebook-gpus/geforce-945m GeForce 945M] * [https://www.geforce.com/hardware/notebook-gpus/geforce-940mx '''GeForce 940MX'''] * [http://www.geforce.com/hardware/notebook-gpus/geforce-940m GeForce 940M] * [https://www.geforce.com/hardware/notebook-gpus/geforce-930mx '''GeForce 930MX'''] * [http://www.geforce.com/hardware/notebook-gpus/geforce-930m GeForce 930M] * [https://www.geforce.com/hardware/notebook-gpus/geforce-920mx '''GeForce 920MX'''] * [http://www.geforce.com/hardware/notebook-gpus/geforce-920m GeForce 920M] * [https://www.geforce.com/hardware/notebook-gpus/geforce-910m GeForce 910M] * [https://www.nvidia.com/coolstuff/demos#!/apollo-11 Apollo 11 Lunar Landing Demo] * [https://developer.nvidia.com/nvidia-nsight-visual-studio-edition Nvidia Nsight] * [http://www.techpowerup.com/gpudb techPowerUp! GPU Database]

{{DEFAULTSORT:GeForce 900 Series}} 900 series Category:Graphics cards Category:False advertising Category:Computer-related introductions in 2014