Volta (microarchitecture)

From Wikipedia, the free encyclopedia
Jump to navigation Jump to search
Nvidia Volta
Release date December 7, 2017
Codename Volta
Transistors 12 nm
Enthusiast Tesla V100
Titan V
Titan V CEO Edition
Quadro GV100
Predecessor Pascal
Successor Turing

Volta, an Nvidia-developed GPU microarchitecture codename, succeeds Pascal and was announced as a future roadmap ambition in March 2013. The architecture is named after Alessandro Volta, the Italian physicist, chemist, and inventor of the electrical battery.

The Nvidia Titan V is the first consumer grade graphics card which uses the Volta microarchitecture. The first graphics card to use it is the Tesla V100, e.g. as part of the Nvidia DGX-1 system.

The Pascal-based GP100 chip was released on 20 June 2016, and the Volta-based GV100 chip was released on 21 June 2017. It is the first chip to feature Tensor cores, specially designed cores that have superior deep learning performance over regular CUDA cores.[1]


In March 2013, Nvidia announced that the successor of the Pascal microarchitecture would be named Volta and include on-chip, stacked DRAM with 1 TB/s of bandwidth.[2][3]

It was reported in March 2017 that TSMC would be fabricating Volta using a 12 nm process and that the new microarchitecture is expected in 2018.[4]

Architectural improvements of the Volta architecture include the following:

  • CUDA Compute Capability 7.0
  • High Bandwidth Memory 2[4][5]
  • NVLink 2.0: a high-bandwidth bus between the CPU and GPU, and between multiple GPUs. Allows much higher transfer speeds than those achievable by using PCI Express; estimated to provide 25 Gbit/s per lane.[6] (Disabled for Titan V)
  • Tensor cores: A tensor core is a unit that multiplies two 4×4 FP16 matrices, and then adds a third FP16 or FP32 matrix to the result by using fused multiply–add operations, and obtains an FP32 result that could be optionally demoted to an FP16 result.[7] Tensor cores are intended to speed up the training of neural networks.[7]
  • PureVideo Feature Set I hardware video decoding
  • Nvidia RTX[8][9]


Volta has been announced as the GPU microarchitecture within the Xavier generation of Tegra SoC focusing on self-driving cars.[10]

Nvidia officially announced the Volta microarchitecture as part of the Tesla V100 product announcement on May 10, 2017. The Volta GV100 GPU is built on a 12 nm process size using HBM2 memory with 900 GB/s of bandwidth.[11]

Nvidia officially announced the NVIDIA TITAN V on December 7, 2017.[12][13]

Nvidia officially announced the Quadro GV100 on March 27, 2018.[14]

Model Launch Code Name (s) Fab
Die size
Bus Interface Core config SM
L2 Cache
Size (MiB)
Clock speeds Fillrate Memory Processing power (GFLOPS) TDP
NVLink 2 Support Launch Price
Base core
clock (MHz)
Boost Clock
Bus width



Nvidia Titan V[15] December 7, 2017 GV100-400-A1 12 21.1 815 PCIe 3.0 ×16 5120:320:96 640 80 6 4.5 1200 1455 1696 139.7? 384 12 652.8 HBM2 3072 13,800 6,900 27,600 250 No $2,999
Nvidia Quadro GV100[16] March 27, 2018 GV100-???-A1 5120:320:128 6 1132 1628 185.2 463 32 870 4096 14,800 7,400 29,600 Yes $8,999
Nvidia Titan V CEO Edition[17][18] N/A GV100-???-A1 1200 1450? 195.8? 489.6? 900? 15,700? 7,800? 31,000? Yes N/A
  1. ^ One Streaming Multiprocessor encompasses 64 CUDA cores and 4 TMUs.
  2. ^ One Graphics Processing Cluster encompasses fourteen Streaming Multiprocessors.
  3. ^ CUDA cores : Texture mapping units : Render output units
  4. ^ A Tensor core is a mixed-precision FPU specifically designed for matrix arithmetic.


Volta is also reported to be included in the Summit and Sierra supercomputers, used for GPGPU compute.[19][20] The Volta GPUs will connect to the POWER9 CPUs via NVLink 2.0, which is expected to support cache coherency and therefore improve GPGPU performance.[21][6][22]

See also[edit]


  1. ^ "NVIDIA Volta AI Architecture | NVIDIA". NVIDIA. Retrieved 2018-04-11. 
  2. ^ Gasior, Geoff (19 March 2013). "Nvidia's Volta GPU to feature on-chip DRAM". The Tech Report. Retrieved 14 March 2017. 
  3. ^ Smith, Ryan (19 March 2013). "Nvidia Updates GPU Roadmap; Announces Volta Family for Beyond 2014". AnandTech. Retrieved 14 March 2017. 
  4. ^ a b Killian, Zak (14 March 2017). "Report: TSMC set to fabricate Volta and Centriq on 12-nm process". The Tech Report. Retrieved 14 March 2017. 
  5. ^ Gasior, Geoff (March 19, 2013). "Nvidia's Volta GPU to feature on-chip DRAM". The Tech Report. 
  6. ^ a b Shah, Agam (22 August 2016). "Nvidia's NVLink 2.0 will first appear in Power9 servers next year". PC World. Retrieved 14 March 2017. 
  7. ^ a b Harris, Mark (May 11, 2017). "CUDA 9 Features Revealed: Volta, Cooperative Groups and More". Retrieved August 12, 2017. 
  8. ^ https://nvidianews.nvidia.com/news/nvidia-rtx-technology-realizes-dream-of-real-time-cinematic-rendering
  9. ^ https://www.geforce.com/whats-new/articles/nvidia-rtx-real-time-game-ray-tracing
  10. ^ Cutress, Ian; Tallis, Billy (4 January 2016). "CES 2017: Nvidia Keynote Liveblog". AnandTech. Retrieved 9 January 2017. 
  11. ^ Smith, Ryan (10 May 2017). "Nvidia Volta Unveiled". AnandTech. Retrieved 2 June 2017. 
  12. ^ https://nvidianews.nvidia.com/news/nvidia-titan-v-transforms-the-pc-into-ai-supercomputer
  13. ^ https://www.nvidia.com/en-us/titan/titan-v/
  14. ^ https://nvidianews.nvidia.com/news/nvidia-reinvents-the-workstation-with-real-time-ray-tracing
  15. ^ "Introducing NVIDIA TITAN V: The World's Most Powerful PC Graphics Card". NVIDIA. Retrieved 2017-12-08. 
  16. ^ "NVIDIA Quadro GV100". Retrieved 2018-03-27. 
  17. ^ Smith, Ryan. "NVIDIA Unveils & Gives Away New Limited Edition 32GB Titan V "CEO Edition"". Retrieved 2018-07-06. 
  18. ^ "NVIDIA TITAN V CEO Edition". TechPowerUp. Retrieved 2018-07-07. 
  19. ^ Shankland, Steven (14 September 2015). "IBM, Nvidia land $325M supercomputer deal". CNET. Retrieved 29 December 2015. 
  20. ^ Noyes, Katherine (16 March 2015). "IBM, Nvidia rev HPC engines in next-gen supercomputer push". PC World. Retrieved 29 December 2015. 
  21. ^ Smith, Ryan (17 November 2014). "Nvidia Volta, IBM Power9 Land Contracts for New US Government Supercomputers". Anandtech. Retrieved 14 March 2017. 
  22. ^ Lilly, Paul (January 25, 2017). "NVIDIA 12nm FinFET Volta GPU Architecture Reportedly Replacing Pascal In 2017". HotHardware. 

External links[edit]