Jump to content

GeForce 10 series: Difference between revisions

From Wikipedia, the free encyclopedia
Browse history interactively
← Previous editNext edit →
Content deleted Content added
VisualWikitext
No edit summary
Tags: Mobile edit Mobile web edit
→External links: added link to cryptocurrencies list, with detailed info on different models of the GTX 1060 cards
Line 501: Line 501:
* [http://developer.nvidia.com/nvidia-nsight-visual-studio-edition Nvidia Nsight]
* [http://developer.nvidia.com/nvidia-nsight-visual-studio-edition Nvidia Nsight]
* [http://www.techpowerup.com/gpudb techPowerUp! GPU Database]
* [http://www.techpowerup.com/gpudb techPowerUp! GPU Database]
* [https://www.pyramidreviews.com/cryptocorner/mining-gpu-speed-and-price-comparison/ethereum-mining-speed-gpu-comparison-for-gtx1060-6gb/ List of GTX 1060s used for Mining Cryptocurrencies]


{{NVIDIA}}
{{NVIDIA}}

Revision as of 23:45, 12 October 2017

GeForce GTX 10 series
The GTX 1070 Founders Edition reference card.
Release dateMay 2016
CodenameGP10x
ArchitecturePascal
ModelsGeForce GTX Series
Transistors
  • 1.8B 14 nm (GP108)
  • 3.3B 14 nm (GP107)
  • 4.4B 16 nm (GP106)
  • 7.2B 16 nm (GP104)
  • 12B 16 nm (GP102)
Cards
Entry-levelGeForce GT 1030
Mid-rangeGeForce GTX 1050
GeForce GTX 1050 Ti
GeForce GTX 1060
High-endGeForce GTX 1070
GeForce GTX 1080
Geforce GTX 1080 Ti
EnthusiastNVIDIA TITAN X
NVIDIA TITAN Xp
API support
DirectXDirect3D 12.0 (feature level 12_1)
OpenCLOpenCL 1.2
OpenGLOpenGL 4.6
VulkanVulkan 1.0
SPIR-V
History
PredecessorGeForce 900 series

Based on Pascal, the GeForce GTX 10 Series is a series of graphics processing units developed by Nvidia, succeeding the GeForce 900 Series.

Architecture

The Pascal microarchitecture, named after Blaise Pascal, was announced in March 2014.[1], and the first Graphics Cards based on it, the GeForce GTX 1080 and 1070 were announced on May 6th, 2016 and released on May 27th, 2016 and June 10th respectively. It incorporates TSMC's 16 nm FinFET technology and succeeds the Maxwell microarchitecture.[2] Samsung Electronics and Nvidia also have an agreement to shrink the design's die to use Samsung's 14 nm FinFET technology.[3]

New Features in GP10x:

  • CUDA Compute Capability 6.0 (GP100 only), 6.1 (GP102, GP104, GP106, GP107, GP108)
  • DisplayPort 1.4
  • HDMI 2.0b
  • Fourth generation Delta Color Compression
  • PureVideo Feature Set H hardware video decoding HEVC Main10 (10 bit), Main12 (12 bit) & VP9 hardware decoding (GM200 & GM204 did not support HEVC Main10/Main12 & VP9 hardware decoding)[4]
  • HDCP 2.2 support for 4K DRM protected content playback & streaming (Maxwell GM200 & GM204 lack HDCP 2.2 support, GM206 supports HDCP 2.2)[5]
  • NVENC HEVC Main10 10 bit hardware encoding
  • GPU Boost 3.0
  • Simultaneous Multi-Projection
  • HB SLI Bridge Technology
  • New memory controller with GDDR5X & GDDR5 support (GP102, GP104) [6]
  • Dynamic load balancing scheduling system.[7] This allows the scheduler to dynamically adjust the amount of the GPU assigned to multiple tasks, ensuring that the GPU remains saturated with work except when there is no more work that can safely be distributed.[7] [7]
  • Instruction-level preemption.[8] In graphics tasks, the driver restricts this to pixel-level preemption because pixel tasks typically finish quickly and the overhead costs of doing pixel-level preemption are much lower than performing instruction-level preemption.[8] Compute tasks get either thread-level or instruction-level preemption.[8] Instruction-level preemption is useful because compute tasks can take long times to finish and there are no guarantees on when a compute task finishes, so the driver enables the very expensive instruction-level preemption for these tasks.[8]
  • Triple buffering implemented in the driver level.[9] Nvidia calls this "Fast Sync".[9] This has the GPU maintain three frame buffers per monitor.[9] This results in the GPU continuously rendering frames, and the most recently completely rendered frame is sent to a monitor each time it needs one.[9] This removes the initial delay that double buffering with vsync causes and disallows tearing.[9] The costs are that more memory is consumed for the buffers and that the GPU will consume power drawing frames that might be wasted because two or more frames could possibly be drawn between the time a monitor is sent a frame and the time the same monitor needs to be sent another frame.[9] In this case, the latest frame is picked, causing frames drawn after the previously displayed frame but before the frame that is picked to be completely wasted.[9] This feature has been backported to Maxwell-based GPUs in driver version 372.70.[10]

Nvidia has announced that the Pascal GP100 GPU will feature four High Bandwidth Memory stacks, allowing a total of 16 GB HBM2 on the highest-end models,[11] 16 nm technology,[2] Unified Memory and NVLink.[12]

Successor architecture

Volta

Main article: Volta (microarchitecture)

After Pascal, the next architecture is codenamed Volta, named after Alessandro Volta. Volta was originally Maxwell's successor, but in 2014 Nvidia announced that Pascal will succeed Maxwell "to take advantage of stacked memory and other innovations sooner."[13] On May 11, 2017, Nvidia announced the first Volta-powered GPU, the Tesla V100 datacenter-targeted card. Built on a 12-nanometer manufacturing process, its GPU measures 815 square mm. It features 21 billion transistors, 5120 CUDA cores running at 1455MHz boost clock, and 640 “tensor cores”, specialized in deep learning. The card also features NVLink 2.0 and 16GB of 4096-bit second-gen HBM2, which peaks to 900 GB/s.[14][15]

Products

An inno3D GeForce GTX 1050 Twin X2.

Founders Edition

Announcing the GeForce 10-series products, Nvidia has introduced Founders Edition graphics card versions of the GTX 1060, 1070, 1080 and 1080 Ti. These are what were previously known as reference cards, i.e. which were designed and built by Nvidia and not by its authorized board partners. These cards have started being used as reference to measure performance of partner cards. The Founders Edition cards have a die cast machine-finished aluminum body with a single radial fan and a vapor chamber cooling (1080, 1080 Ti only[16]), an upgraded power supply and a new low profile backplate (1070, 1080, 1080 Ti only). [17] Nvidia also released a limited supply of Founders Edition cards for the GTX 1060 that were only available directly from Nvidia's website.[18] Founders Edition cards prices (with the exception of the GTX 1080 Ti) are greater than MSRP of partners cards, however some partners' cards, incorporating a complex design, with liquid or hybrid cooling may cost more than Founders Edition.

GeForce 10 (10xx) series

Model Launch Code name(s) Fab (nm) Transistors (billion) Die size (mm2) Bus interface Core config[b] SM Count [c] Clock speeds Fillrate Memory Processing power (GFLOPS)[d] TDP (watts) SLI HB support[e] Launch price (USD)
Base core clock (MHz) Boost core clock (MHz) Memory (MT/s) Pixel (GP/s)[f] Texture (GT/s)[g] Size (GiB) Bandwidth (GB/s) Bus type Bus width (bit) Single precision (Boost) Double precision (Boost) Half precision (boost)[20] MSRP Founders Edition
GeForce GT 1030[21] May 17, 2017 GP108-300-A1 14 1.8[22] 70 PCIe 3.0 x4 [23][24] 384:24:16 3 1227 1468 6000 19.6 29.4 2 48 GDDR5 64 942 (1127) 29 (35) 15 (18) 30 No $80[25] —
GeForce GTX 1050[26] October 25, 2016 GP107-300-A1 3.3 132 [27] PCIe 3.0 x16 640:40:32 5 [28] 1354 1455 7000 43.3 54.2 112 128 1733 (1862) 54 (58) 27 (29) 75 $109
GeForce GTX 1050 Ti[26] GP107-400-A1 768:48:32 6 [29] 1290 1392 41.3 61.9 4 1981 (2138) 62 (67) 31 (33) $139
GeForce GTX 1060 3GB[30] August 18, 2016 GP106-300-A1 16 4.4 200 [31] 1152:72:48 9 [32] 1506 1708 8000
9000[h] 		72.3 108.4 3 192
216[h] 		192 3470 (3935) 108 (123) 54 (61) 120 $199
GeForce GTX 1060 6GB[30] July 19, 2016 GP106-400-A1
GP106-410-A1[h] 		1280:80:48 10 [33] 120.5 6 3855 (4372) 120 (137) 60 (68) $249 $299
GeForce GTX 1070[34] June 10, 2016 GP104-200-A1 7.2 314 [35] 1920:120:64 15 [36] 1683 8000 96.4[i][37] 180.7 8 256 256 5783 (6463) 181 (202) 90 (101) 150 2-way SLI HB[38] or traditional 2/3/4-way SLI[39] $379 $449
GeForce GTX 1080[19] May 27, 2016 GP104-400-A1
GP104-410-A1[h] 		2560:160:64 20 [40] 1607 1733 10000
11000[h] 		102.8 257.1 320
352[h] 		GDDR5X 8228 (8873) 257 (277) 128 (139) 180 $599 $699
GeForce GTX 1080 Ti[41] March 10, 2017 GP102-350-K1-A1 12 471 3584:224:88 28 [42] 1480 1582 11000 130.2 331.5 11 484 352 10609 (11340) 332 (354) 166 (177) 250 $699 $699
NVIDIA TITAN X[43] August 2, 2016 GP102-400-A1 3584:224:96 1417 1531 10000 136 317.4 12 480 384 10157 (10974) 317 (343) 159 (171) 250 — $1200
NVIDIA TITAN Xp[44] April 6, 2017 GP102-450-A1 3840:240:96 30 1405[45] 1582 11410 152 379.7 547.7 384 10790 (12150) ? (380) ? (190) 250 — $1200
  1. ^ The NVIDIA TITAN Xp & the Founders Edition GTX 1080 Ti does not have a dual link DVI port, but a DisplayPort to single link DVI adapter is included in the box.
  2. ^ Shader Processors : Texture mapping units : Render output units
  3. ^ The number of streaming multiprocessors on the GPU.
  4. ^ For calculating the processing power, see the Performance subsection of the Pascal architecture article.
  5. ^ SLI HB only supports a maximum of 2-way SLI using SLI HB bridges, however if using traditional SLI bridges it can support a maximum of 4-way SLI but the performance is mostly improved in synthetic benchmarks only.
  6. ^ 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.
  7. ^ Texture fillrate is calculated as the number of TMUs multiplied by the base core clock speed.
  8. ^ a b c d e f GTX 1060 and GTX 1080 cards shipped after April 2017 feature increased memory speeds, thus increasing memory bandwidth.
  9. ^ The GTX 1070 has one of the four GPCs disabled in the die. Losing one of the Raster Engines only allows for the use of 48 ROPs per cycle.

GeForce 10 (10xx) series for notebooks

The biggest highlight to this line of notebook GPUs is the implementation of configured specifications close to (for the GTX 1060–1080) and exceeding (for the GTX 1050/1050 Ti) that of their desktop counterparts, as opposed to having "cut-down" specifications in previous generations. As a result, the "M" suffix is completely removed from the model's naming schemes, denoting these notebook GPUs to possess similar performance to those made for desktop PCs, including the ability to overclock their core frequencies by the user, something not possible with previous generations of notebook GPUs. This was made possible having lower TDP ratings as compared to their desktop equivalents, making these desktop-level GPUs thermally feasible to be implemented into OEM notebook chassis with improved thermal dissipation designs, and as such are only available through the OEMs that. In addition, the GP104 and GP106-based Notebook GPUs also come in lower-TDP and quieter variations called the Max-Q Design specifically made for ultra-thin gaming systems in conjunction with OEM Partners that incorporate enhanced heat dissipation mechanisms with lower operating noise volumes, largely made as an additional more powerful option to existing gaming notebooks as well, which was launched on 27th June 2017.

Model Launch Code name Fab (nm) Transistors (billion) Die size (mm2) Bus interface Core config SM Count Clock speeds Fillrate Memory API support (version) Processing power (GFLOPS) TDP (watts) SLI support
Base core clock (MHz) Boost core clock (MHz) Memory (MT/s) Pixel (GP/s) Texture (GT/s) Size (GiB) Bandwidth (GB/s) Bus type Bus width (bit) DirectX OpenGL OpenCL Vulkan Single precision (Boost) Double precision Half precision
GeForce MX150 (GT 1030)[46] May 25, 2017 GP108-300 14 1.8 74 PCIe 3.0 x4 384:24:8[47] 3 1468 1532 6000 11.7 35.2 2 48 GDDR5 64 12.0 (12_1) 4.6 1.2 1.0 1127 (1177) ? ? 25 No
GeForce GTX 1050 (Notebook)[48] January 3, 2017 GP107-300 (N17P-G0) 3.3 135 PCIe 3.0 x16 640:40:16 5 1354 1493 7000 43.3 54.2 Up to 4 112 128 1733 (1911) 27 14 75
GeForce GTX 1050 Ti (Notebook)[48] GP107-400 (N17P-G1) 768:48:32 6 1493 1620 47.8 71.7 2293 (2488) 36 18
GeForce GTX 1060 (Notebook)[48] August 16, 2016 GP106-400 (N17E-G1) 16 4.4 200 1280:80:48 10 1404 1670 8000 67.4 112 Up to 6 192 192 3594 (4275) 112 56 80
GeForce GTX 1070 (Notebook)[48] GP104-200 (N17E-G2) 7.2 314 2048:128:64 16 1442 1645 92.3 185 8 256 256 5906 (6738) 185 92 110 Yes
GeForce GTX 1080 (Notebook)[48] GP104-400 (N17E-G3) 2560:160:64 20 1556 1733 10000 99.6 249 320 GDDR5X 7967 (8873) 249 124 150

Chipset table

Main article: Comparison table of GeForce 10 Series

See also

References

  1. ^ "NVIDIA Updates GPU Roadmap; Announces Pascal". The Official NVIDIA Blog.
  2. ^ a b "Talks of foundry partnership between NVIDIA and Samsung (14nm) didn't succeed, and the GPU maker decided to revert to TSMC's 16nm process". Retrieved August 25, 2015.
  3. ^ "Samsung to Optical-Shrink NVIDIA "Pascal" to 14 nm". Retrieved August 13, 2016.
  4. ^ "How The New Pascal Architecture Supports Next-Generation Video Playback". May 17, 2016. Retrieved June 11, 2016.
  5. ^ "Nvidia Pascal HDCP 2.2". Nvidia Hardware Page. Retrieved May 8, 2016.
  6. ^ Shrout, Ryan (July 14, 2016). "3DMark Time Spy: Looking at DX12 Asynchronous Compute Performance". PC Perspective. Retrieved July 14, 2016.
  7. ^ a b c Smith, Ryan (July 20, 2016). "The NVIDIA GeForce GTX 1080 & GTX 1070 Founders Editions Review: Kicking Off the FinFET Generation". AnandTech. p. 9. Retrieved July 21, 2016.
  8. ^ a b c d Smith, Ryan (July 20, 2016). "The NVIDIA GeForce GTX 1080 & GTX 1070 Founders Editions Review: Kicking Off the FinFET Generation". AnandTech. p. 10. Retrieved July 21, 2016.
  9. ^ a b c d e f g Smith, Ryan; Wilson, Derek (July 20, 2016). "The NVIDIA GeForce GTX 1080 & GTX 1070 Founders Editions Review: Kicking Off the FinFET Generation". AnandTech. p. 13. Retrieved July 22, 2016.
  10. ^ "Release 370 Graphics Drivers for Windows, Version 372.70" (PDF). August 30, 2016. Retrieved August 30, 2016.
  11. ^ Harris, Mark (April 5, 2016). "Inside Pascal: NVIDIA's Newest Computing Platform". Parallel Forall. Nvidia. Retrieved June 3, 2016.
  12. ^ "NVIDIA Pascal GPU Architecture to Provide 10X Speedup for Deep Learning Apps - NVIDIA Blog". The Official NVIDIA Blog.
  13. ^ "Nvidia's Pascal to use stacked memory, proprietary NVLink interconnect". The Tech Report. Retrieved March 26, 2014.
  14. ^ "Nvidia's bleeding-edge Volta GPU: 5 things PC gamers need to know". PCWorld.
  15. ^ "Nvidia's monstrous Volta GPU appears, packed with 21 billion transistors and 5,120 cores". PCWorld.
  16. ^ Hruska, Joel (May 31, 2016). "Nvidia's new GTX 1070 freezes out Fury, trashes Titan X". ExtremeTech. Retrieved June 5, 2016.
  17. ^ Burnes, Andrew (May 18, 2016). "GeForce GTX 1080 Founders Edition: Premium Construction & Advanced Features". GeForce.com. Retrieved May 19, 2016.
  18. ^ "A Quantum Leap for Every Gamer: NVIDIA Unveils the GeForce GTX 1060". Nvidia Newsroom. July 7, 2016. Retrieved October 18, 2016. {{cite web}}: Cite has empty unknown parameter: |1= (help)
  19. ^ a b Nvidia. "GTX 1080 Graphics Card". Retrieved May 7, 2016.
  20. ^ Smith, Ryan (July 20, 2016). "FP16 Throughput on GP104: Good for Compatibility (and Not Much Else) - The NVIDIA GeForce GTX 1080 & GTX 1070 Founders Editions Review: Kicking Off the FinFET Generation". AnandTech. Retrieved July 22, 2016.
  21. ^ Nvidia. "GeForce GT 1030 | GeForce". Retrieved May 17, 2017.
  22. ^ Angelini, Chris (13 July 2017). "Nvidia GeForce GT 1030 2GB Review". Tom's Hardware. Retrieved 13 July 2017.
  23. ^ http://www.palit.com/palit/vgapro.php?id=2883&lang=en&pn=NE5103000646-1080F&tab=sp
  24. ^ https://www.msi.com/Graphics-card/GeForce-GT-1030-AERO-ITX-2G-OC.html#hero-specification
  25. ^ Paul, Ian (17 May 2017). "Nvidia quietly launches the GeForce GT 1030, a Radeon RX 550 rival with a modest price". Retrieved 17 May 2017.
  26. ^ a b Nvidia. "GTX 1050 Graphics Card". Retrieved October 18, 2016.
  27. ^ Hagedoorn, Hilbert. "Palit GeForce GTX 1050 Ti KalmX Review - Introduction". Retrieved February 24, 2017.
  28. ^ "NVIDIA GeForce GTX 1050". techpowerup.com.
  29. ^ "NVIDIA GeForce GTX 1050 Ti". techpowerup.com.
  30. ^ a b Nvidia. "GTX 1060 Graphics Card". Retrieved August 18, 2016.
  31. ^ Angelini, Chris. "Nvidia GeForce GTX 1060 6GB Review". Retrieved July 19, 2016.
  32. ^ "NVIDIA GeForce GTX 1060 3 GB". techpowerup.com.
  33. ^ "NVIDIA GeForce GTX 1060 6 GB". techpowerup.com.
  34. ^ Nvidia. "GTX 1070 Graphics Card". Retrieved May 18, 2016.
  35. ^ Shrout, Ryan. "The GeForce GTX 1080 8GB Founders Edition Review - GP104 Brings Pascal to Gamers". Retrieved May 17, 2016.
  36. ^ "NVIDIA GeForce GTX 1070". techpowerup.com.
  37. ^ Smith, Ryan. "Synthetics - The NVIDIA GeForce GTX 1080 & GTX 1070 Founders Editions Review: Kicking Off the FinFET Generation". Retrieved 2016-07-21.
  38. ^ W1zzard (May 17, 2016). "Nvidia GeForce GTX 1080 8 GB". TechPowerUp. Retrieved May 17, 2016.{{cite web}}: CS1 maint: numeric names: authors list (link)
  39. ^ W1zzard (June 21, 2016). "Nvidia GeForce GTX 1080 SLI". TechPowerUp. Retrieved June 21, 2016.{{cite web}}: CS1 maint: numeric names: authors list (link)
  40. ^ "NVIDIA GeForce GTX 1080". techpowerup.com.
  41. ^ https://www.nvidia.com/en-us/geforce/products/10series/geforce-gtx-1080-ti/
  42. ^ "NVIDIA TITAN X Pascal". techpowerup.com.
  43. ^ Nvidia. "NVIDIA TITAN X Graphics Card with Pascal". Retrieved July 21, 2016.
  44. ^ Nvidia. "TITAN Xp Graphics Card with Pascal Architecture". Retrieved April 6, 2017.
  45. ^ "NVIDIA TITAN Xp". TechPowerUp.
  46. ^ "Introducing GeForce MX150 Laptops: Supercharged For Work and Play". Retrieved 2017-07-17.
  47. ^ "NVIDIA GeForce MX150". TechPowerUp. Retrieved 2017-07-17.
  48. ^ a b c d e "GeForce GTX 10-Series Notebooks". geforce.com.
Retrieved from "https://en.wikipedia.org/w/index.php?title=GeForce_10_series&oldid=805085972"