NVIDIA B100 vs AMD Radeon RX 470D
We compared two Desktop platform GPUs: 96GB VRAM B100 and 4GB VRAM Radeon RX 470D to see which GPU has better performance in key specifications, benchmark tests, power consumption, etc.
Main Differences
NVIDIA B100 's Advantages
Released 8 years and 1 months late
Boost Clock has increased by 52% (1837MHz vs 1206MHz)
More VRAM (96GB vs 4GB)
Larger VRAM bandwidth (4.10TB/s vs 211.2GB/s)
15104 additional rendering cores
AMD Radeon RX 470D 's Advantages
Lower TDP (120W vs 1000W)
Score
Benchmark
FP32 (float)
B100
+1336%
62.08 TFLOPS
Radeon RX 470D
4.322 TFLOPS
Graphics Card
Nov 2024
Release Date
Oct 2016
Server Blackwell
Generation
Arctic Islands
Desktop
Type
Desktop
PCIe 5.0 x16
Bus Interface
PCIe 3.0 x16
Clock Speeds
1665 MHz
Base Clock
926 MHz
1837 MHz
Boost Clock
1206 MHz
2000 MHz
Memory Clock
1650 MHz
Memory
96GB
Memory Size
4GB
HBM3e
Memory Type
GDDR5
4096bit
Memory Bus
256bit
4.10TB/s
Bandwidth
211.2GB/s
Render Config
-
Compute Units
28
132
SM Count
-
16896
Shading Units
1792
528
TMUs
112
24
ROPs
32
528
Tensor Cores
-
-
-
-
256 KB (per SM)
L1 Cache
16 KB (per CU)
50 MB
L2 Cache
2 MB
-
-
-
Theoretical Performance
44.09 GPixel/s
Pixel Rate
38.59 GPixel/s
969.9 GTexel/s
Texture Rate
135.1 GTexel/s
248.3 TFLOPS
FP16 (half)
4.322 TFLOPS
62.08 TFLOPS
FP32 (float)
4.322 TFLOPS
31.04 TFLOPS
FP64 (double)
270.1 GFLOPS
Board Design
1000W
TDP
120W
1400 W
Suggested PSU
300 W
No outputs
Outputs
1x HDMI 2.0b
3x DisplayPort 1.4a
-
Power Connectors
1x 6-pin
Graphics Processor
GB102
GPU Name
Ellesmere
-
GPU Variant
Polaris 10 PROD
Blackwell
Architecture
GCN 4.0
TSMC
Foundry
GlobalFoundries
5 nm
Process Size
14 nm
104 billion
Transistors
5.7 billion
Unknown
Die Size
232 mm²
Graphics Features
N/A
DirectX
12 (12_0)
N/A
OpenGL
4.6
3.0
OpenCL
2.1
N/A
Vulkan
1.2
10.1
CUDA
-
N/A
Shader Model
6.4
