Last week we had the opportunity to attend an online event where AMD shared with us the keys to the new AMD Radeon Pro W7900 and Radeon Pro W7800, two high-performance professional cards that are based on the RDNA3 architecture, and that have Specific optimizations for the professional sector.
In case someone has been lost, I remind you that this range competes with NVIDIA’s A-series RTXformerly known as NVIDIA Quadro, and have nothing to do with the Radeon Instinct graphics accelerators, since these do not have image output and are aimed at servers and data centers, while the Radeon Pro do have image output and are They are mainly aimed at workstations and very diverse sectors, such as content creation, design and manufacturing, architecture and construction, among others.
The Radeon Pro W7900 and Radeon Pro W7800 start from the same design that we have seen in the Radeon RX 7000, which means that they use a chiplet type design at the GPU level, although this one is very particular, because in reality the graphics core itself uses a single packageknown as GDC, and is surrounded by six MCD chiplets that hold the infinite cache, which is nothing more than L3 cache interconnected with the GDC.
With the RDNA3 architecture, AMD has achieved a significant leap in terms of performance, and has also introduced important changes that must be highlighted to better understand the evolution of the Radeon Pro W7900 and Radeon Pro W7800. This new architecture is optimized for professional workflows focused on design, video editing, and multitasking and the GPU is made on TSMC’s 5nm node (cache chiplets use 6nm node).
The computing unit, which is the basis of the RDNA3 architecture, has 64 dual broadcast stream processorsdispose of two units specialized in accelerating tasks related to artificial intelligence and has a second generation ray tracing acceleration core which, according to AMD, improves performance by up to 50% compared to first-generation models.
AMD has also improved the multimedia engine, which translates into higher performance in encoding and decoding tasks, supports encode and decode up to 8K at 60 FPS with the new AV1 codecsuses artificial intelligence to improve the process and supports the new connectivity standard Display Port 2.1, that allows you to use displays with 8K resolution and 120 Hz. According to AMD, the Radeon Pro W7900 lalmost double the performance of the Radeon Pro W6800and the company has also highlighted its value in relation to price-performance.
Specifications and price of the Radeon Pro W7900
- Navi 31 graphics core with Multi-chiplet Die design.
- GPU made in 5nm and L3 cache made in 6nm.
- 96 computing units.
- 6,144 shaders.
- 384 texturing units.
- 192 raster units.
- 96 units for second generation ray tracing acceleration.
- 192 units for artificial intelligence acceleration.
- GPU at 2,269 MHz-2,499 MHz, normal and turbo mode.
- 61.42 TFLOPs of power in FP32.
- 384-bit bus.
- 48 GB of GDDR6 memory with ECC technology and a bandwidth of 864 GB/s.
- 96 MB of infinite cache with a bandwidth of 5.3 TB/s.
- 295 watt TBP.
- Price: $3,999.
Specifications and price of the Radeon Pro W7800
- Navi 31 graphics core with Multi-chiplet Die design.
- GPU made in 5nm and L3 cache made in 6nm.
- 70 computing units.
- 4,480 shaders.
- 280 texturing units.
- 128 raster units.
- 70 units for second generation ray tracing acceleration.
- 140 units for artificial intelligence acceleration.
- 2 MHz GPU, normal and turbo mode.
- 41 TFLOPs of power in FP32.
- 256-bit bus.
- 32 GB of GDDR6 memory with ECC technology and a bandwidth of 576 GB/s.
- 64 MB of infinite cache with a bandwidth of 5.3 TB/s.
- 260 watt TBP.
- Price: $2,499.
