Yesterday Apple introduced
The youngest in the new series was the Radeon Pro 5300 video card, which is built on the Navi GPU with only 20 Compute units (CUs) and, accordingly, 1280 stream processors. The GPU clock speed is not specified, but the peak performance value is 4,2 Tflops (FP32). The amount of RAM GDDR6 is 4 GB.
Up a notch is the Radeon Pro 5500 XT, which has 8GB of memory and a 24 CU GPU with 1536 stream processors. Its performance level is 5,3 teraflops, which is 0,1 teraflops faster than the consumer Radeon RX 5500 XT. Next comes the Radeon Pro 5700, which is built on a chip with 36 CU, that is, with 2304 stream processors and has 8 GB of GDDR6. Here, the performance level is 6,2 teraflops, which is much lower than the performance of the Radeon RX 5700, which offers 7,95 teraflops.
Finally, the oldest new product is the Radeon Pro 5700 XT video card. It offers 2560 stream processors and performance up to 7,6 teraflops. For comparison, the consumer Radeon RX 5700 XT is capable of delivering 9,75 teraflops. Apparently, such a big difference is due to the fact that for use in cramped iMac conditions, the new product is significantly limited in terms of power consumption, and, accordingly, in terms of frequencies. Curiously, the new Radeon Pro 5700 XT has 16 GB of GDDR6 memory instead of 8 GB in the publicly available Radeon RX 5700 XT.
As you can see from the slides above, the new graphics cards offer a significant performance boost over the AMD Vega-based solutions that have been used in the past.
Source:
Source: 3dnews.ru