Apr 2, 2014
The next generation GPU, NVIDIA Pascal, is all out- smarter, sleeker, smaller, and more efficient. This device is sure to change the way we handle data and move it around. The speed at which data is exchanged between the CPU and GPU usually limits the processor’s ability to perform. NVIDIA strives to solve this problem by introducing NVLink, a chip-to-chip communication technology whose speed will definitely be impressed. Data processing is one of the integral facets of the modern-day economy. Therefore, people would need their data processed as fast as possible.
NVLink is obviously designed to speed up the GPU and CPU communication, it will also improve data transfer between GPUs. This means that even graphic transfers will be improved greatly. This is speed, power and efficiency put together. The graphic transfer has been the headache for many people who usually use graphics in their places of work. With NVLink, you can be rest assured that your data are on the right track, quite literally.
The speed, power and efficiency of the GPU come at a cost of a big chip. NVIDIA says it’s working on that. Actually, the GPU is already the biggest chip in the world. Its interface is already extremely wide. The solution here is to build chips on top of other chips. And NVIDIA will build smaller GPUs that are four times as energy efficient as today’s chips. Being small in size and efficiency is very important when it comes data transfer chips. Many people need chips to fit into very small devices.
The NVIDIA Company is not satisfied with the achievements it has already made. It is working on other technologies to improve its other forthcoming GPU. This includes working on a unified memory system that will eliminate the need for developers to split resources between CPU and GPU memory. Thus, a Pascal-powered laptop is a dream that will soon come true, thanks to NVIDIA.
This next-generation GPU, code-named Pascal, features stacked 3D DRAM, unified memory, and a technology called NVLink to bring more power to today’s small and thin devices.
Pascal has three main features:
Stacked DRAM or 3D memory is where chips are stacked It allows the DRAM to be in the same package as the GPU. And the whole thing is designed to enable the GPUs fast access to the data held in RAM. It will definitely boost efficiency and output.
The more compact and smaller GPUs can also be fitted into smaller devices. This offers a several-fold increase in bandwidth, more than twice the memory capacity and quadrupled energy efficiency.
The unified memory will allow GPU applications and CPU applications to access each other’s memory. This will save resources that would have been allocated between the two chips.
NVLink greatly boosts the memory capacity between the CPU and GPU from the current 16GB/s to more than 80GB/s. This will ensure efficiency and fastness in accessing data stored in the two chips. What more could somebody interested in faster data transfer ask for?