AMD Big Navi Graphics Exposure: FP32 Calculates up to 18T, or this Friday

AMD is scheduled to hold a conference of financial analysts at its California headquarters on March 6 from 5 a.m. to 10 a.m. Beijing time, where CEO Dr. Su Zifeng is rumored to be announcing The Radeon “Big Navi” graphics card. The “Big Navi” is said to be the flagship of three new graphics products, with single-precision (FP32) floating-point performance of up to 18TFLOPS, which is 80% higher than the RX 5700 XT or RTX 2080, and 30% higher than the Radeon VII or RTX 2080 Ti.

If previous revelations are to be believed, the Big Navi or the RX 5950 XT, the other two may be the RX 5900 XT and the RX 5800 XT.

According to the RX 5950 XT has 80 sets of Cu (twice as many as the RX 5700 XT), the FP32 can also double, the core main frequency tuning, the strength of 18TFLOPS. However, this also brings a few questions, such as whether it will be a “glue” double core, and how to suppress power consumption (5700 XT TDP is 225 to 235W), the same 7nm and 3840 stream processor Radeon VII power consumption has touched 300 watts.

AMD Big Navi Graphics Exposure: FP32 Calculates up to 18T, or this Friday

Of course, 18T is nothing compared to the exposed NVIDIA Ampere core “Fat Man” and “Little Boy.” Of the 7552 and 6912 CUDAs, the FP32 computing power is as high as 33.5T and 27.9T, respectively.