We have new information on NVIDIA next-Gen architecture Pascal, which involves the NVIDIA GeForce GTX 1080. It is expected that in GTC 2016, the CEO of NVIDIA, Jen-Hsun will officially announce the new Pascal graphics card, but a reliable source, indicate that the GeForce GTX 1080 would launch in May and would not feature on the HBM2 memory, but its silicon Pascal GP104 is manufactured under 16nm FinFET featuring amazing performance and power efficiency. The GeForce GTX 1080 would be accompanied by 8GB of GDDR5X memory and the GPU will be fed through a single 8-Pin connector (TDP 150-175W).
NVIDIA Pascal Based GeForce GTX 1080 with 8GB GDDR5X Memory Coming in May
The source also indicates that the HBM2 memory would be exclusively for the high-end GP100 (Full Pascal Core) GPU core. The GP100 consumer variant will feature up to 16 GB HBM2 memory while the HPC variants will ship with up to 32 GB HBM2 memory. Hopefully pretty high for those who want access to that memory prices.
The GDDR5X memory was one of the most interesting as was already mentioned in its day options, it should be much more affordable than HBM2 memory addition, to reach a speed around 10 and 14 Gbps, but its peak performance is fixed in the 16 Gbps, almost double if you compare it with GDDR5 memory, but still far from the 1024 Gbps that can be achieved through HBM2 memory.
In terms of VRAM, the GDDR5X memory can also provide increased capacity, while the GDDR5 memory can be found in configurations of 512 MB, 1GB, 2GB, 4GB and 8GB, the GDDR5X memory can be found in capacities of 4GB, 6GB, 12GB and maximum 16GB. With respect to the HBM2 memory, NVIDIA is currently in cooperation with Samsung Electronics, while AMD is with SK Hynix.
Previously at the GTC March, 2015, we heard a lot from NVIDIA about Pascal based architecture, but a few changes and improvements are noticed as compared with previous one.
- Pascal micro architecture.
- DirectX 12 feature level 12_1 or higher.
- Successor to the GM200 GPU found in the GTX Titan X and GTX 980 Ti.
- Built on the 16FF+ manufacturing process from TSMC.
- Allegedly has a total of 17 billion transistors, more than twice that of GM200.
- Features NVLink and support for Mixed Precision FP16 compute tasks at twice the rate of FP32 and full FP64 support. 2016 release.