The new Pascal graphics processing unit (GPU) architecture was revealed by NVIDIA last month, and since then, it has captured the imagination of gamers who do not mind spending more to get the best gaming experience. However, focus is now shifting to the NVIDIA Pascal GP100 GPU as benchmark results of it have appeared online.

Reporting on the benchmarks for the NVIDIA Pascal GP100 GPU which is being used on the Tesla P100 graphics board, WCCFTech states that the benchmarks were achieved by using a tool named Assisted Model Building with Energy Refinement (AMBER). This particular tool was used recently to see how the GP100 GPU stands up against other GPUs manufactured by NVIDIA and not those created by other competitors.

More importantly, when AMBER was used, a GP100 GPU was pitted against other GPUs from NVIDIA, which includes the GP104 and GK110. As a result of this benchmark test, it was revealed that a Tesla P100 was able to do better than a quad Titan X configuration. Also, when comparing the GeForce GTX 1080 against the GP100 GPU, it was found to be equally fast. Given these results, it is estimated that the Tesla P100 is faster and it will result in hyperscale servers performing much better.

As of now, WCCFTech states that the Tesla P100 graphics board is already in the process of shipping out to customers who will welcome the benchmark results. The Tesla P100 GPU was announced by NVIDIA about two months ago, and since then, it has been shipping out to server partners of the company. Announced by Jen-Hsun Huang, who is the CEO of NVIDIA, it arrived during the opening keynote of GTC 2016. The Tesla P100 since then has been produced in a large volume and after it begins shipping this month, it will be employed by deep learning/cloud computing servers. Also designed for use in supercomputers, it will become available through OEMs in 2017.