GeForce 700 series

From Wikipedia, the free encyclopedia
Jump to: navigation, search
GeForce 700 Series
Codename GF117, GF119, GK104, GK106, GK107, GK110, GK208, GM107
Models GeForce Series
  • GeForce GT Series
  • GeForce GTX Series
Fabrication process and transistors
  • 585M 28 nm (GF117)
  • 1,270M 28 nm (GK107)
  • Unknown 28 nm (GK208)
  • 3,540M 28 nm (GK104)
  • 7,080M 28 nm (GK110)
Cards
Entry-level GeForce GT 730
GeForce GT 740
Mid-range GeForce GTX 750
GeForce GTX 750 Ti
GeForce GTX 760
GeForce GTX 760 Ti
High-end GeForce GTX 770
GeForce GTX 780
Enthusiast GeForce GTX 780 Ti
GeForce GTX Titan
GeForce GTX Titan Black
GeForce GTX Titan Z
Rendering support
Direct3D full Direct3D 11.0 and partial Direct3D 12.0 support[1]
OpenCL OpenCL 1.1
OpenGL OpenGL 4.4 in production drivers.
OpenGL 4.5 in beta drivers.[2]
History
Predecessor GeForce 600 series
Variant GeForce 800M series
Successor GeForce 900 series

The GeForce 700 Series is a family of graphics processing units developed by Nvidia, used in desktop and laptop PCs. It is mainly based on a refresh of the Kepler microarchitecture (GK-codenamed chips) used in the previous GeForce 600 Series, but also includes cards based on the previous Fermi (GF) and later Maxwell (GM) architectures. A number of GeForce 700 series chips were released for mobile devices in April 2013. GeForce 700 series cards were first released in May 2013, starting with the release of the GeForce GTX Titan on February 19, 2013, and the GeForce GTX 780 on May 23, 2013.

Overview

GK110 has been designed and is being marketed with compute performance in mind. It contains 7.1 billion transistors. This model also attempts to maximise energy efficiency through the execution of as many tasks as possible in parallel according to the capabilities of its streaming processors.

With GK110, increases in memory space and bandwidth for both the register file and the L2 cache over previous models, are seen. At the SMX level, GK110's register file space has increased to 256KB composed of 65K 32bit registers, as compared to Fermi's 33K 32bit registers totaling 128 KB. As for the L2 cache, GK110 L2 cache space increased by up to 1.5MB, 2x as big as GF110. Both the L2 cache and register file bandwidth have also doubled. Performance in register-starved scenarios is also improved as there are more registers available to each thread. This goes in hand with an increase of total number of registers each thread can address, moving from 63 registers per thread to 255 registers per thread with GK110.

With GK110, Nvidia also reworked the GPU texture cache to be used for compute. With 48KB in size, in compute the texture cache becomes a read-only cache, specializing in unaligned memory access workloads. Furthermore error detection capabilities have been added to make it safer for use with workloads that rely on ECC.[3]

This series will support DirectX 12.[4]

Architecture

A Gigabyte GTX770 graphics card
The PCB of a GTX 780

The GeForce 700 Series contains features from both GK104 and GK110. Kepler based members of the 700 series add the following standard features to the GeForce family.

Derived from GK104 :

  • DisplayPort 1.2
  • HDMI 1.4a 4K x 2K video output
  • Purevideo VP5 hardware video acceleration (up to 4K x 2K H.264 decode)
  • Hardware H.264 encoding acceleration block (NVENC)
  • Support for up to 4 independent 2D displays, or 3 stereoscopic/3D displays (NV Surround)
  • Bindless Textures
  • GPU Boost
  • TXAA
  • Manufactured by TSMC on a 28 nm process

New Features from GK110 :

  • Compute Focus SMX Improvement
  • CUDA Compute Capability 3.5
  • New Shuffle Instructions
  • Dynamic Parallelism
  • Hyper-Q (Hyper-Q's MPI functionality reserve for Tesla only)
  • Grid Management Unit
  • NVIDIA GPUDirect (GPU Direct’s RDMA functionality reserve for Tesla only)

Compute Focus SMX Improvement

With GK110, Nvidia opted to increase compute performance. The single biggest change from GK104 is that rather than 8 dedicated FP64 CUDA cores, GK110 has up to 64, giving it 8x the FP64 throughput of a GK104 SMX. The SMX also sees an increase in space for register file. Register file space has increased to 256KB compared to Fermi. The texture cache are also improved. With a 48KB space, the texture cache can become a read-only cache for compute workloads.[3]

New Shuffle Instructions

At a low level, GK110 sees additional instructions and operations to further improve performance. New shuffle instructions allow for threads within a warp to share data without going back to memory, making the process much quicker than the previous load/share/store method. Atomic operations are also overhauled, speeding up the execution speed of atomic operations and adding some FP64 operations that were previously only available for FP32 data.[3]

NVENC

Main article: Nvidia NVENC

Hyper-Q

Hyper-Q expands GK110 hardware work queues from 1 to 32. The significance of this being that having a single work queue meant that Fermi could be under occupied at times as there wasn’t enough work in that queue to fill every SM. By having 32 work queues, GK110 can in many scenarios, achieve higher utilization by being able to put different task streams on what would otherwise be an idle SMX. The simple nature of Hyper-Q is further reinforced by the fact that it’s easily map to MPI, a common message passing interface frequently used in HPC. As legacy MPI-based algorithms that were originally designed for multi-CPU systems that became bottlenecked by false dependencies now have a solution. By increasing the number of MPI jobs, it’s possible to utilize Hyper-Q on these algorithms to improve the efficiency all without changing the code itself.[3]

Microsoft DirectX Support

NVIDIA Kepler GPUs of the GeForce 700 series fully support DirectX 11.0.

NVIDIA will partially support the DX12 API on all the DX11-class GPUs it has shipped; these belong to the Fermi, Kepler and Maxwell architectural families.[1]

Dynamic Parallelism

Dynamic Parallelism ability is for kernels to be able to dispatch other kernels. With Fermi, only the CPU could dispatch a kernel, which incurs a certain amount of overhead by having to communicate back to the CPU. By giving kernels the ability to dispatch their own child kernels, GK110 can both save time by not having to go back to the CPU, and in the process free up the CPU to work on other tasks.[3]

Products

GeForce 700 (7xx) series

The GeForce 700 series for desktop architecture. Cheaper and lower performing products were expected to be released over time. Kepler supports 11.1 features with 11_0 feature level through the DirectX 11.1 API, however Nvidia did not enable four non-gaming features in Hardware in Kepler (for 11_1).[5][6]

  • 1 Shader Processors : Texture mapping units : Render output units
  • 2 Pixel fillrate is calculated as the number of ROPs multiplied by the base core clock speed
  • 3 Texture fillrate is calculated as the number of TMUs multiplied by the base core clock speed.
  • 4 Single precision performance is calculated as 2 times the number of shaders multiplied by the base core clock speed.
  • 5 Double precision performance of the GTX Titan & GTX Titan Black is either 1/3 or 1/24 of single-precision performance depending on a user-selected configuration option in the driver that boosts single-precision performance if double-precision is set to 1/24 of single-precision performance,[7] while other Kepler chips' double precision performance is fixed at 1/24 of single-precision performance.[8] GeForce 700 series Maxwell chips' double precision performance is 1/32 of single-precision performance.[9]
  • 6 SLI support connecting up to 4 identical GPUs card for a 4-way SLI configuration. Those support 4-way SLI can support 3-way & 2-way SLI, however a Dual-GPUs card is already 2-way SLI configuration internally therefore they support 4-way SLI with an identical Dual-GPUs card but do not support 3-way SLI.
Model Launch Code name Fab (nm) Transistors (Million) Die size (mm) GPU count Bus interface Memory (MiB) Core config1 Clock speeds Fillrate Memory API support (version) Processing Power (GFLOPS) GFLOPS/W Single Precision TDP (watts) SLI support6 Release Price (USD)
Base core clock (MHz) Boost core clock (MHz) Memory (MT/s) Pixel (GP/s)2 Texture (GT/s)3 Bandwidth (GB/s) Bus type Bus width (bit) DirectX OpenGL OpenCL Single precision4 Double precision5
GeForce GT 720 [10] Mar 27, 2014 GK208 28 1020 79 1 PCIe 2.0 x8 1024
2048
192:16:8 797 N/A 1800
5000
6.4 12.8 14.4
40
DDR3
GDDR5
64 11.0 4.5 1.1 306 12.8 16.1 19 N/A 49
GeForce GT 730 (DDR3, 128-bit)[11] Jun 18, 2014 GF108 40 585 116 1 PCIe 2.0 x16 1024 96:16:4 700 N/A 1800 2.8 11.2 28.8 DDR3 128 11.0 4.5 1.1 134 Unknown 2.7 49 N/A Unknown
GeForce GT 730 (DDR3, 64-bit)[11] Jun 18, 2014 GK208 28 1020 79 1 PCIe 2.0 x16 2048 384:16:8 902 N/A 1800 7.2 14.4 14.4 DDR3 64 11.0 4.5 1.1 693 28.9 30.1 23 N/A Unknown
GeForce GT 730 (GDDR5)[11] Jun 18, 2014 GK208 28 1020 79 1 PCIe 2.0 x16 1024 384:16:8 902 N/A 5000 7.2 14.4 40 GDDR5 64 11.0 4.5 1.1 693 28.9 27.7 25 N/A Unknown
GeForce GT 740 (DDR3)[12] May 29, 2014 GK107 28 1270 118 1 PCIe 3.0 x16 2048 384:32:16 993 N/A 1800 15.9 31.8 28.8 DDR3 128 11.0 4.5 1.1 762 31.8 11.9 64 N/A $89
GeForce GT 740 (GDDR5)[12] May 29, 2014 GK107 28 1270 118 1 PCIe 3.0 x16 1024 384:32:16 993 N/A 5000 15.9 31.8 80 GDDR5 128 11.0 4.5 1.1 762 31.8 11.9 64 N/A $89
GeForce GTX 745 (OEM)[13] Feb 18, 2014 GM107 28 1870 148 1 PCIe 3.0 x16 4096 384:24:16 1033 Unknown 1800 16.5 24.8 28.8 DDR3 128 11.2 4.5 1.1 793 24.8 14.4 55 N/A OEM
GeForce GTX 750 [14] February 29, 2014 GM107 28 1870 148 1 PCIe 3.0 x16 1024 512:32:16 1020 1085 5000 16.3 32.6 80 GDDR5 128 11.2 4.5 1.1 1044 32.6 19 55 N/A $119
GeForce GTX 750 Ti [15] February 18, 2014 GM107 28 1870 148 1 PCIe 3.0 x16 1024
2048
640:40:16 1020 1085 5400 16.3 40.8 86.4 GDDR5 128 11.2 4.5 1.1 1306 40.8 21.8 60 N/A $149
GeForce GTX 760 192-bit [16] Unknown GK104 28 3540 294 1 PCIe 3.0 x16 1536
3072
1152:96:24 823 888 5808 19.8 79 134 GDDR5 192 11.0 4.5 1.1 1896 79 14.6 130 3-way OEM
GeForce GTX 760 [17] June 25, 2013 GK104 28 3540 294 1 PCIe 3.0 x16 2048
4096
1152:96:32 980 1033 6008 31.4 94.1 192 GDDR5 256 11.0 4.5 1.1 2258 94.1 13.3 170 3-way $249
GeForce GTX 760 Ti [18] Unknown GK104 28 3540 294 1 PCIe 3.0 x16 2048 1344:112:32 915 980 6008 29.3 103 192 GDDR5 256 11.0 4.5 1.1 2460 103 14.5 170 3-way OEM
GeForce GTX 770 [19] May 30, 2013 GK104 28 3540 294 1 PCIe 3.0 x16 2048

4096

1536:128:32 1046 1085 7008 33.5 134 224 GDDR5 256 11.0 4.5 1.1 3213 134 14.0 230 3-way $399[20]
GeForce GTX 780 [21] May 23, 2013 GK110 28 7080 561 1 PCIe 3.0 x16 3072 2304:192:48 863 900 6008 41.4 166 288 GDDR5 384 11.0 4.5 1.1 3977 166 15.9 250 3-way $649[20]
GeForce GTX 780 Ti [22] November 7, 2013 GK110 28 7080 561 1 PCIe 3.0 x16 3072 2880:240:48 876 928 7000 42.0 210 336 GDDR5 384 11.0 4.5 1.1 5046 210 20.2 250 4-way $699[20]
GeForce GTX Titan [23] February 19, 2013 GK110 28 7080 561 1 PCIe 3.0 x16 6144 2688:224:48 837 876 6008 40.2 188 288 GDDR5 384 11.0 4.5 1.1 4500 1500 18.0 250 4-way $999
GeForce GTX Titan Black [24] February 18, 2014 GK110 28 7080 561 1 PCIe 3.0 x16 6144 2880:240:48 889 980 7000 42.7 213 336 GDDR5 384 11.0 4.5 1.1 5121 1707 20.5 250 4-way $999
GeForce GTX Titan Z[25] March 25, 2014 2× GK110 28 2× 7080 2× 561 2 PCIe 3.0 x16 2× 6144 2× 2880:240:48 705 876 7000 2× 33.8 2× 169 2× 336 GDDR5 2× 384 11.0 4.5 1.1 8122 2707 21.7 375 4-way on 2 cards $2999

GeForce 700M (7xxM) series

Some implementations may use different specifications.

Model Launch Code name Fab (nm) Bus interface Memory (MiB) Core config1 Clock speed Fillrate Memory API support (version) Processing Power2
(GFLOPS)
TDP (watts) Notes
Core (MHz) Shader (MHz) Memory (MT/s) Pixel (GP/s) Texture (GT/s) Bandwidth (GB/s) Bus type Bus width (bit) DirectX OpenGL OpenCL
GeForce 705M [26] June 1, 2013 GF119 40 PCIe 2.0 x16 up to 2048 48:8:4 775 1550 1800  ?  ?  ? DDR3 64 11.0 4.1 1.1  ? 12W Rebadged 520M
GeForce 710M [27] April 1, 2013 GF117 28 PCIe 2.0 x16 up to 2048 96:16:4 775 1550 1800  ?  ? 14.4 DDR3 64 11.0 4.4 1.1  ?  ?
GeForce GT 720M [28] April 1, 2013 GF117//GK208 28 PCIe 2.0 x16 up to 2048 96:16:4/192:16:8 800 1600 2000  ?  ? 16.0 DDR3 64 11.0 4.4 1.1  ?  ?
GeForce GT 730M [29] April 1, 2013 GK107/GK208 28 PCIe 3.0 x16/2.0 x8 up to 4096 384:32:16(8) 725 725 1800 - 4000  ?  ? 14.4 - 64.0 DDR3/GDDR5 64/128 11.0 4.4 1.1  ?  ?
GeForce GT 735M [30] April 1, 2013 GK208 28 PCIe 2.0 x8 up to 2048 384:32:8 889 889 2000  ?  ? 16.0 DDR3 64 11.0 4.4 1.1  ?  ?
GeForce GT 740M [31] April 1, 2013 GK107/GK208 28 PCIe 3.0 x16/2.0 x8 up to 2048 384:32:16(8) 810/1033 810/1033 1800/3600  ?  ? 14.4 - 57.6 DDR3/GDDR5 128/64 11.0 4.4 1.1  ?  ?
GeForce GT 745M [32] April 1, 2013 GK107 28 PCIe 3.0 x16 up to 2048 384:32:16 837 837 2000 - 5000  ?  ? 32.0 - 80.0 DDR3/GDDR5 128 11.0 4.4 1.1  ?  ?
GeForce GT 750M [33] April 1, 2013 GK107 28 PCIe 3.0 x16 up to 4096 384:32:16 967 967 2000 - 5000  ?  ? 32 - 80 DDR3/GDDR5 128 11.0 4.4 1.2 722.7  ?
GeForce GT 755M [34] Unknown GK107 28 PCIe 3.0 x16 up to 2048 384:32:16 980? 980? 5400  ?  ? 86.4 GDDR5 128 11.0 4.4 1.1  ?  ?
GeForce GTX 760M [35] May 30, 2013 GK106 28 PCIe 3.0 x16 2048 768:64:16 657 657 4008  ?  ? 64.1 GDDR5 128 11.0 4.4 1.1  ?  ?
GeForce GTX 765M [36] May 30, 2013 GK106 28 PCIe 3.0 x16 2048 768:64:16 850 850 4008  ?  ? 64.1 GDDR5 128 11.0 4.4 1.1  ?  ?
GeForce GTX 770M [37] May 30, 2013 GK106 28 PCIe 3.0 x16 3072 960:80:24 811 811 4008  ?  ? 96.2 GDDR5 192 11.0 4.4 1.1  ?  ?
GeForce GTX 780M [38] May 30, 2013 GK104 28 PCIe 3.0 x16 4096 1536:128:32 823 823 5000  ?  ? 160.0 GDDR5 256 11.0 4.4 1.1  ?  ?

Chipset table

See also

References

  1. ^ a b Kowaliski, Cyril (March 21, 2014). "DirectX 12 will also add new features for next-gen GPUs". The Tech Report. Retrieved April 1, 2014. 
  2. ^ https://developer.nvidia.com/opengl-driver
  3. ^ a b c d e "NVIDIA Launches Tesla K20 & K20X: GK110 Arrives At Last". AnandTech. 11/12/2012. 
  4. ^ http://blogs.nvidia.com/blog/2014/03/20/directx-12/
  5. ^ NVIDIA Kepler not fully compliant with DirectX 11.1
  6. ^ Nvidia Doesn't Fully Support DirectX 11.1 with Kepler GPUs, But… - Bright Side Of News
  7. ^ GK110 The True Tank - Nvidia GeForce GTX Titan 6 GB GK110 On A Gaming Card
  8. ^ Nvidia GeForce GTX 780 Ti Review GK110, Fully Unlocked - GK110, Unleashed The Wonders Of Tight Binning
  9. ^ Smith, Ryan; T S, Ganesh (February 18, 2014). "The NVIDIA GeForce GTX 750 Ti and GTX 750 Review: Maxwell Makes Its Move". AnandTech. p. 5. Retrieved February 18, 2014. 
  10. ^ GeForce GT 720 | Specifications | GeForce
  11. ^ a b c GeForce GT 730 | Specifications | GeForce
  12. ^ a b GeForce GT 740 | Specifications | GeForce
  13. ^ GeForce GTX 745 (OEM) | Specifications | GeForce
  14. ^ GeForce GTX 750 | Specifications | GeForce
  15. ^ GeForce GTX 750 Ti | Specifications | GeForce
  16. ^ GeForce GTX 760 192-bit | Specifications | GeForce
  17. ^ GeForce GTX 760 | Specifications | GeForce
  18. ^ GeForce GTX 760 Ti | Specifications | GeForce
  19. ^ GeForce GTX 770 | Specifications | GeForce
  20. ^ a b c http://www.bit-tech.net/news/hardware/2013/10/28/nvidia-geforce-gtx-780-ti-price-and-release/1
  21. ^ GeForce GTX 780 | Specifications | GeForce
  22. ^ GeForce GTX 780 Ti | Specifications | GeForce
  23. ^ GeForce GTX TITAN | Specifications | GeForce
  24. ^ GeForce GTX Titan Black | Specifications | GeForce
  25. ^ Walker, Justin (25 March 2014). "Two GPUs, One Insane Graphics Card: Introducing the GeForce GTX TITAN Z". Nvidia Blog. Nvidia Corporation. Retrieved 25 March 2014. 
  26. ^ GeForce 705M | Specifications | GeForce
  27. ^ GeForce 710M | Specifications | GeForce
  28. ^ GeForce GT 720M | Specifications | GeForce
  29. ^ GeForce GT 730M | Specifications | GeForce
  30. ^ GeForce GT 735M | Specifications | GeForce
  31. ^ GeForce GT 740M | Specifications | GeForce
  32. ^ GeForce GT 745M | Specifications | GeForce
  33. ^ GeForce GT 750M | Specifications | GeForce
  34. ^ GeForce GT 755M | Specifications | GeForce
  35. ^ GeForce GTX 760M | Specifications | GeForce
  36. ^ GeForce GTX 765M | Specifications | GeForce
  37. ^ GeForce GTX 770M | Specifications | GeForce
  38. ^ GeForce GTX 780M | Specifications | GeForce

External links