Radeon Pro Duo Polaris vs GeForce RTX 5090
We compared two Desktop platform GPUs: 16GB VRAM Radeon Pro Duo Polaris and 32GB VRAM GeForce RTX 5090 to see which GPU has better performance in key specifications, benchmark tests, power consumption, etc.
Main Differences
Radeon Pro Duo Polaris 's Advantages
Lower TDP (250W vs 575W)
GeForce RTX 5090 's Advantages
Released 7 years and 9 months late
Boost Clock2407MHz
More VRAM (32GB vs 16GB)
Larger VRAM bandwidth (1.79TB/s vs 224.0GB/s)
19456 additional rendering cores
Score
Benchmark
FP32 (float)
Radeon Pro Duo Polaris
5728
GeForce RTX 5090
+1729%
104800
Graphics Card
Apr 2017
Release Date
Jan 2025
Radeon Pro GCN
Generation
GeForce 50
Desktop
Type
Desktop
PCIe 3.0 x16
Bus Interface
PCIe 5.0 x16
Clock Speeds
-
Base Clock
2017 MHz
-
Boost Clock
2407 MHz
1750 MHz
Memory Clock
1750 MHz
Memory
16GB
Memory Size
32GB
GDDR5
Memory Type
GDDR7
256bit
Memory Bus
512bit
224.0GB/s
Bandwidth
1.79TB/s
Render Config
36
Compute Units
-
-
SM Count
170
2304
Shading Units
21760
144
TMUs
680
32
ROPs
176
-
Tensor Cores
680
-
RT Cores
170
16 KB (per CU)
L1 Cache
128 KB (per SM)
2 MB
L2 Cache
96 MB
Theoretical Performance
39.78 GPixel/s
Pixel Rate
423.6 GPixel/s
179.0 GTexel/s
Texture Rate
1637 GTexel/s
5.728 TFLOPS
FP16 (half)
104.8 TFLOPS
5.728 TFLOPS
FP32 (float)
104.8 TFLOPS
358.0 GFLOPS
FP64 (double)
1.637 TFLOPS
Board Design
250W
TDP
575W
600 W
Suggested PSU
950 W
1x HDMI 2.0b
3x DisplayPort 1.4a
Outputs
1x HDMI 2.1b
3x DisplayPort 2.1b
1x 6-pin + 1x 8-pin
Power Connectors
1x 16-pin
Graphics Processor
Ellesmere
GPU Name
GB202
Ellesmere Gemini GL
GPU Variant
GB202-300-A1
GCN 4.0
Architecture
Blackwell 2.0
GlobalFoundries
Foundry
TSMC
14 nm
Process Size
5 nm
5.7 billion
Transistors
92.2 billion
232 mm²
Die Size
750 mm²
Graphics Features
12 (12_0)
DirectX
12 Ultimate (12_2)
4.6
OpenGL
4.6
2.1
OpenCL
3.0
1.3
Vulkan
1.4
-
CUDA
12.0
6.7
Shader Model
6.8
