NVIDIA’s next-gen Volta GPUs will be monstrous, with three brand-new high-end graphics cards on the method. NVIDIA has dominated the high-end graphics card market for a while now. The company will surely be the forefront with the release of a new graphic architecture.
Moving beyond 2017, Volta will replace the Pascal refresh in 2018. GDDR6 chips rated at 16Gbps are expected to enhance memory bandwidth by 60 percent. This is in comparison to the GDDR5X 10Gbps memory technology, which was deployed in the GTX 1080 and Titan X.
According to WCCFTECH, the ‘professional version’ of Volta-based graphics card will be powered by HBM2. The company did not clarify the ‘customer’ or ‘enthusiasts’ variation. It is still uncertain whether the variation will be powered by either the HBM2 or GDDR6.
NVIDIA Volta GPU Launch Set in 2017
The target launch for the Volta-based graphics card is in 2017. These GPU cards are for professional use such as innovative servers and the likes. On the other hand, NVIDIA is targeting to release the ‘consumer’ Volta versions next year.
Incidentally, GDDR6 will likewise appear throughout the very same timeline. PC World reported that the GDDR6 will be a quicker version with as much as 14Gbps (bits per second). The report said that the GDDR6 will be more power-efficient compared to GDDR5. This is similar with the HBM technology, which used greater performance and less power usage.
The Summit Supercomputer has 5-10x improvement in application efficiency, increasing the Titan supercomputer that featured the Kepler GK110 GPU architecture. The Titan is consisted of 18,688 nodes ranked at 1.4 TF, per node. The Summit Supercomputer features around 4,600 nodes with a ranked compute output of over 40 TF, per node.
Each node has the two IBM Power9 CPUs and six NVIDIA Volta V100 GPUs. NVIDIA’s NVLINK2 adjoined will be totally incorporated between these nodes. The system will consume 13 MW peak power that is simply 4 MW more than the Titan supercomputer, which is 9 MW. NVIDIA previously stated that the Volta GV100 GPUs will deliver an accuracy floating General Matrix Multiply of 72 GFLOPS. The 72 GFLOPS will deliver more compared to the 42 GFLOPS on Pascal GP100.
NVIDIA Volta GPU Will Have Two Variations
Using the mentioned provision, a Volta GV100 with a TDP of 300W can provide 9.5 TFLOPs of double precision performance. This is almost two times that of the present generation GP100 GPU. The final configuration of Volta V100 and IBM Power9 CPUs will make the Summit Supercomputer as the leading machine. Its efficiency crossed the 250 Petaflops mark which is impressive for experts.
The chances of seeing the NVIDIA Volta GPU powered by HBM2 might show up throughout the latter part of 2017. It will supposedly reveal the ‘expert’ or ‘enterprise’ variations. On the other hand, the ‘enthusiasts’ variation will come in 2018.