New GPU Memory Standards to Supercharge Graphics
Virtual Reality and the ever increasing interest in life-like computer game graphics has led to the development of very advanced graphics for computers and mobile devices. With the development of these new technologies in mind, companies such as Intel, Nvidia and AMD are designing chips that work better with higher quality graphics. Improvements to chip design are being made, such as developing faster ways in which Graphic Processing Units can create and broadcast images.
To accomplish this feat companies are also developing units with faster Graphic Processing Unit architecture and increasing the speed of computer memory is another way in which computer graphics are being improved.
The GDDR5X and HBM2 memory standards are examples of improvements in memory that have recently been developed. Both of these standards were recently approved by the JEDEC Solid State Technology Association and their effectiveness in producing faster memory operations is already being honed by companies such as Samsung, Micron and AMD.
HBM2 is considered revolutionary primarily due to the changes in the way it helps move data in Graphic Processing Units. Unlike other memory standards HBM2 cells are stacked in a 3D format, thereby making the movement of data within its cells much faster. As stated by Nathan Brookwood, principal analyst at Insight 64,
“With HBM2, there are multiple chips stacked and instead of increasing the frequency, you are widening the interface.”
In a similar manner GDDR5X is twice as fast as its predecessor, GDDR5. It is an incremental improvement, with a memory structure that has been redefined and has data rates of 10 to 14 gigabits per second. Both of these memory standards no doubt will have huge impact on how computer graphics are developed and how Graphic Processing Units function.