In the CPU field, AMD has won a game. In terms of GPU, AMD and old opponent NVIDIA still have a big gap. It is mainly reflected in the market share. The GPU technology has its own strengths. Product technology architect SAM Naffziger has talked about the frequency advantage in an interview.
The entire interview of SAM Naffziger is mainly about the efficiency of the graphics card. In terms of power management, it is inevitable to involve the frequency -related part. Everyone knows that the higher the frequency, the higher the power consumption, but AMD is optimized in various ways. It can achieve high frequencies without adding complex gate circuits.
According to him,AMD's graphics card can currently achieve the frequency of 2.5GHz, which is a unique advantage of AMD and can maintain high energy.
As for friendly business NVIDIA, Sam Naffziger believes that their GPU frequency can also achieve the frequency of 2.5GHz or more, for exampleHowever, this will increase the voltage to a very high level, such as 1.2V or 1.3V, andThe power consumption is directly proportional to the secondary square of the voltage, which will bring higher power consumption, and AMD is achieved under the appropriate voltage.
Sam Naffziger's AMD graphics card high frequency is indeed an advantage, but what he did not mention is actually AMD's advantage in craftsmanship compared to NVIDIA.2.5GHz above high frequency has a lot of credit for TSMC 7nm process, NVIDIA's RTX 30 graphics card is also a Samsung 8nm process. Based on the Samsung 10nm improvement, there is a difference from one to two generations as 7nm.
AMD and NVIDIA's new generation of products this year will use TSMC's 5nm process.2.5GHz is already the basic disk, and it is normal to get up to 3GHz.At this time, we can really see the design level of the two GPUs.