The Nvidia Pascal GPU is said to see the light of day in April. The latest and most powerful graphics processing unit from the Santa Clara-based manufacturer will allegedly be unveiled at the upcoming GPU Technology Conference.
Nvidia's CEO and co-founder Jen-Hsun Huang is believed to host a keynote during the event on April 5, where he is expected to lay out the company's plans for its Pascal graphics cards and ultimately reveal the latest in the powerful family of GPUs.
Rather than bring a prototype, Huang is allegedly showcasing a working unit, which means that more specific details about the performance of the Nvidia Pascal GPU will be revealed during the affair.
With its launch set for April, the Nvidia Pascal GPU is speculated to hit the market by June. According to reports, the Computex conference in Taipei, Taiwan will mark the start of the gizmo's availability. The first lineup is said to be used in gaming notebooks.
However, the graphics card being based on the 16nm FinFET manufacturing process from TSMC could lead to a delay. Rumor has it that Nvidia is having trouble pulling this off, which could ultimately delay the supposed availability of the Nvidia Pascal GPU in June.
Not too long ago, Nvidia Pascal GPUs were spotted being shipped from Nvidia's various facilities. They carry the serial number 699, which happens to be the same number in the title of Huang's keynote address at the GPU Technology Conference, further hinting at the April reveal.
As for the specs of the Nvidia Pascal GPU, the general consensus is that they will be boasting twice the performance per watt compared to Maxwell. It is also expected to support DirectX 12 feature level 12_1 or higher.
The unit is also expected t feature a 4096-bit memory bus interface as well as compatibility with NV Link. The Nvidia Pascal GPU apparently supports half precision GeForce FP16. It will also purportedly contain 17 billion transistors, which is twice the number of those in the Nvidia Maxwell GM200.
