NVIDIA announced its Pascal architecture based GTX 1000 Series and a new datacenter GPU, namely the Tesla P100 a while ago. The Tesla P100, codenamed GP100, is the most complex and powerful GPU ever made by NVIDIA and this will make way for new GPUs for PC gaming in future.
Currently, four Pascal based GPUs are designed and are going through testing and validation work in NVIDIA’s testing facility in India. NVIDIA is wasting no time in getting its GP100 and GP104 GPUs ready, to be released in the market in the second half of 2016.
So, for those of you who are reading about the Pascal architecture and NVIDIA GTX 1000 Series for the first time, this article will provide you with all the necessary details and information related to both and you will be able to make a better decision and strategy of when and what graphics card should you upgrade to. TwentyNext has written a pretty decent piece comparing the GTX 1000 Series with the with the current GTX 900 series.
What’s new in Nvidia Pascal GPUs
NVIDIA will be unveiling their new “Pascal” architecture GPUs under NVIDIA GTX 1000 Series for both consumers and industrial computing at Computex 2016. The Pascal based GPUs will be the successor and a huge improvement to the Maxwell based GPUs by providing more processing power on smaller chips. The Pascal architecture makes use of High Memory Bandwidth (HMB) and also acts as a bridge between the Maxwell architecture series and the Volta architecture, which NVIDIA is currently developing and is expected to be unveiled in 2018.The Pascal based GPUs will be the successor and a huge improvement to the Maxwell based GPUs
Following is a list of things we know about the Pascal architecture and GP100, the flagship GPU which will utilize this architecture:
- The graphics cards utilizing this architecture will launch in the second half of 2016
- 2 times more performance per watt than Maxwell architecture based GPUs
- GP100 is the successor to the GM200 GPU found in GTX 980 Ti and GTX Titan X
- Support for DirectX 12
- Reported to have a total of 17 billion transistors, more than twice that of GM200
- Contains the NVIDIA NVLink technology that allows faster data transfer between the CPU and GPU
- NVIDIA states that it is 5 to 12 times faster than traditional PCIE 3.0
- Will be built on 16nm FinFET manufacturing process from TSMC
- Features a 4096-bit memory bus interface
NVIDIA was going to release the Volta architecture based GPUs as the successor to the Maxwell architecture. But most of the technologies like Hybrid Memory Cube, a stacked memory standard, for the Volta architecture were not completely developed quickly as NVIDIA had hoped. So Pascal was created which uses High Memory Bandwidth, another stacked memory standard, to serve as a bridge between Maxwell and Volta architectures.
Nvidia GTX 1000 GPU Series
Along with the announcement of the Pascal architecture, NVIDIA will be releasing the GTX 1000 series GPUs in the second half of this year. Not officially announced yet, but NVIDIA plans to announce GeForce GTX 1070 and GeForce GTX 1080 graphics cards at Computex 2016 that starts on the 31st of May 2016.
The GTX 1070 and GTX 1080 cards will contain 8 GB DDR5 memory, 256-bit memory bus and memory speeds of over 8 Gbps and 12 Gbps respectively. Pascal architecture allows lesser number of CUDA cores than in GTX 980, but with more throughput. These cards will make use of the GP104 graphics core which similar to the GP100 graphics core.
The most interesting thing about the availability of these cards is that an initial wave of these cards will be through NVIDIA’s Store before being available in custom design from NVIDIA’s partners like MSI, Gigabyte and ASUS.
This was pretty much everything you need to know about NVIDIA Pascal architecture and the GTX 1000 series graphic cards. The Pascal architecture will be most important for VR, 4K video gaming and DirectX 12 since VR headsets require rendering at 90 FPS and many AAA titles cannot be played at 60FPS in 2K or 4K resolutions. The NVIDIA Pascal architecture GPUs will deliver a lot of processing power while also being energy efficient.
See the Video from GTC 2016
Guest post by: Adeel Inam
Adeel Inam is the editor at TwentyNext where their experienced team provides solutions to various technical problems and cover all the latest technological trends of the 21st century with a focus on VR.