NVIDIA RTX PRO 6000 Blackwell vs NVIDIA B200 SXM6

GPU · Specs Compare

Products
GPU NVIDIA RTX PRO 6000 Blackwell NVIDIA · 2025-03
GPU NVIDIA B200 SXM6 NVIDIA · 2024-11
Graphics Card
Release Date
Mar 2025
Nov 2024
Generation
Blackwell PRO W
Server Blackwell
Type
Desktop
Desktop
Bus Interface
PCIe 5.0 x16
PCIe 6.0 x16
Clock Speeds
Base Clock
1590 MHz
120 MHz
Boost Clock
2617 MHz
1830 MHz
Memory Clock
1750 MHz
2000 MHz
Memory
Memory Size
96GB
180GB
Memory Type
GDDR7
HBM3e
Memory Bus
512bit
8192bit
Bandwidth
1.79TB/s
8.19TB/s
Render Config
SM Count
188
148
Shading Units
24064
18944
TMUs
752
592
ROPs
192
24
Tensor Cores
752
592
RT Cores
188
—
L1 Cache
128 KB (per SM)
256 KB (per SM)
L2 Cache
128 MB
126 MB
Theoretical Performance
Pixel Rate
502.5 GPixel/s
43.92 GPixel/s
Texture Rate
1968.0 GTexel/s
1083.4 GTexel/s
FP16 (half)
126.0 TFLOPS
1109.4 TFLOPS
FP32 (float)
126.0 TFLOPS
69.34 TFLOPS
FP64 (double)
1.968 TFLOPS
34.67 TFLOPS
Graphics Processor
GPU Name
GB202
GB100
Architecture
Blackwell 2.0
Blackwell
Foundry
TSMC
TSMC
Process Size
5 nm
5 nm
Transistors
92.2 billion
208 billion
Die Size
750 mm²
1628 mm²
GPU Variant
—
GB100-886N-A1
Board Design
TDP
600W
1000W
Suggested PSU
1000 W
1400 W
Outputs
4x DisplayPort 2.1b
No outputs
Power Connectors
1x 16-pin
—
Graphics Features
DirectX
12 Ultimate (12_2)
N/A
OpenGL
4.6
N/A
OpenCL
3.0
3.0
Vulkan
1.4
N/A
CUDA
12.0
10.0
Shader Model
6.9
N/A
Benchmark
3DMark Time Spy
42,269
—
3DMark Time Spy Extreme
27,891
—
3DMark Speed Way
15,642
—
▲ Specs Compare Data sourced from official specifications and benchmark tests.