The Radeon PRO W7800 is a professional graphics card by AMD, launched on May 26th, 2023. Built on the 5 nm process, and based on the Navi 31 graphics processor, the card supports DirectX 12 Ultimate. The Navi 31 graphics processor is a very large chip with a die area of 529 mm² and 57.7 billion transistors. Unlike the fully unlocked Radeon RX 7900 XTX, which uses the same GPU but has all 6144 shaders enabled, AMD has disabled some shading units on the Radeon PRO W7800 to reach the product's target shader count. It features 4480 shading units, 280 texture mapping units, and 128 ROPs. Also included are 140 Matrix Cores which help improve the speed of machine learning applications with support for the INT4 and INT8 data formats. The GPU also has 70 raytracing acceleration cores. AMD has paired 32 GB GDDR6 memory with the Radeon PRO W7800, which is connected using a 256-bit memory interface. The GPU is operating at a frequency of 1895 MHz, which can be boosted up to 2525 MHz, memory is running at 2250 MHz (18 Gbps effective). Being a dual-slot card, the AMD Radeon PRO W7800 draws power from 2x 8-pin power connectors, with power draw rated at 260 W maximum. Display outputs include: 3x DisplayPort 2.1, 1x mini-DisplayPort 2.1. Radeon PRO W7800 is connected to the rest of the system using a PCI-Express 4.0 x16 interface. The card's dimensions are 280 mm x 110 mm x 40 mm, and it features a dual-slot cooling solution. Its price at launch was 2,499 US Dollars.
Based on TPU review data: "Performance Summary" at 1920x1080, 4K for RTX 3080 and faster.
Performance estimated based on architecture, shader count and clocks.
Clock Speeds
Base Clock
1895 MHz
Boost Clock
2525 MHz
Memory Clock
2250 MHz
18 Gbps effective
Memory
Memory Size
32 GB
Memory Type
GDDR6
Memory Bus
256 bit
Bandwidth
576.0 GB/s
Render Config
Shading Units
4480
TMUs
280
ROPs
128
Compute Units
70
Matrix Cores
140
RT Cores
70
L0 Cache
32 KB per CU
L1 Cache
256 KB per Array
L2 Cache
6 MB
L3 Cache
64 MB
Theoretical Performance
Pixel Rate
323.2 GPixel/s
Texture Rate
707.0 GTexel/s
FP16
45.25 TFLOPS (1:1)
FP32
45.25 TFLOPS
FP64
1.414 TFLOPS (1:32)
Matrix Performance
INT4
181.0 TOPS
INT8
90.50 TOPS
FP16
90.50 TFLOPS
BF16
90.50 TFLOPS
Numeric Format Support
Vector
INT4
INT8
INT16
INT32
FP16
BF16
FP32
FP64
Matrix
INT1
INT2
INT4
INT8
MXFP4
NVFP4
MXFP6
FP8
MXFP8
FP16
BF16
FP32
TF32
FP64
Board Design
Slot Width
Dual-slot
Length
280 mm
11 inches
Height
110 mm
4.3 inches
Width
40 mm
1.6 inches
TDP
260 W
Suggested PSU
600 W
Outputs
3x DisplayPort 2.1 1x mini-DisplayPort 2.1
Power Connectors
2x 8-pin
Graphics Features
DirectX
12 Ultimate (12_2)
OpenGL
4.6
OpenCL
2.0
Vulkan
1.4
Shader Model
6.9
Navi 31 GPU Notes
Super Resolution: FSR 4.1.1, XeSS 3
Frame Generation: FSR 3.1.6, XeSS 3
Ray Regeneration: No Support
Compute Units: 3rd Gen
Ray Tracing: 2nd Gen
Infinity Cache: 2nd Gen
AI Accelerators: 1st Gen
Radiance Display Engine: 1st Gen
Shader ISA: GFX11.0 (gfx1100)
Graphics/Compute: 11.0.0
Display Core Next: 3.2.0
Video Core Next: 4.0.0
JPEG Engine: 4.0.0
Per SIMD32:
- 10 KB Scalar Register File
Per CU:
- 16 KB Vector Data Cache
- 64 KB Local Data Share (GCN-Mode)
Per WGP:
- 16 KB Scalar Data Cache / K Cache
- 32 KB Instruction Cache
- 128 KB Local Data Share