AMD’s Next 7nm Vega 20 GPU Could Possibly Be A Tesla V100 Rival Not GTX 1180
The forthcoming GPUs of AMD, which is the new 7 nm Vega 20, is forecasted to reach the 20 TFLOP limit. The primary drawback of the said GPU is its large consumption of electric power at 400 W. Reports have it that the latest Radeon Instinct video cards, which is the 7 nm Vega 20 GPUs, are likely to be broadcast on the news in the near future.
A preview of the forthcoming 7nm server parts has already been given by AMD. According to what has been disclosed, ASCII.jp wrapped-up that the results based on the Vega 20 have the capacity to deal out a processing power of 20 TFLOPs.
ASCII.jps assesses the Vega 20 die size at 360 square millimeters, which is a 70 percent cutback in the whole die size as compared to the Vega 10 chip 7 in comparison to the Vega 10 chip. The 7nm Vega 20 GPUs could possibly be the key solution that can rival the TESLA V100 from Nvidia, not the anticipated 1180 GPUs.
Should the reports and estimation be right, the AMD 7nm Vega 20 GPU, which is estimated to showcase 20 TFLOPS of Compute, 32 GB of HBM2 VRAM that is up to 400W TDP. Although this is not traditional or standard one for a gaming platform.
With the 7nm LPP clock rate, the user can either gain a power savings of 55% from equivalent frequencies of 14nm; or a 40% faster clock rate that can likewise be derived from a 14nm power. But if the user’s goal is to attain or have a 40% faster clock with the same bundle of power, the power that will be used is also needed to be transported to the other HBM2 VRAM (4 Stacks vs 2 Stacks) as per the new set of DNN instruction.
So, what are your thoughts on the AMD’s upcoming GPU? Do you think it can compete with Tesla V100? Share your thoughts in the comment section below.