GPU Comparison
GEFORCE
NVIDIA B200
CORE STATE
GB100
VRAM
90 GB
CLOCK SPEED
1965 MHz
TDP
1000 W
BUS WIDTH
4096 bit
ARCHITECTURE
Blackwell
PROCESS
5 nm
LAUNCH DATE
—
VS
GEFORCE
Tesla T4
CORE STATE
TU104
VRAM
16 GB
CLOCK SPEED
1590 MHz
TDP
70 W
BUS WIDTH
256 bit
ARCHITECTURE
Turing
PROCESS
12 nm
LAUNCH DATE
2018
PERFORMANCE BENCHMARKS
geekbench_opencl
geekbench_vulkan
DETAILED SPECIFICATIONS
SPECIFICATION
B200
Tesla T4
Core Specs
Shading Units
18,944
2,560
-86.5%
Shaders
18,944
2,560
-86.5%
TMUs
592
160
-73.0%
ROPs
24
64
+166.7%
SM Count
148
40
-73.0%
Clocks
Base Clock
700 MHz
585 MHz
Boost Clock
1965 MHz
1590 MHz
Memory Clock
2000 MHz
8 Gbps effective
1250 MHz
10 Gbps effective
Memory
Memory Size
90 GB
16 GB
VRAM (MB)
92,160
16,384
-82.2%
Memory Type
HBM3e
GDDR6
Memory Bus
4096 bit
256 bit
Bandwidth
4.10 TB/s
320.0 GB/s
Cache
L1 Cache
256 KB (per SM)
64 KB (per SM)
L2 Cache
50 MB
4 MB
Performance
Pixel Rate
47.16 GPixel/s
101.8 GPixel/s
Texture Rate
1,163.3 GTexel/s
254.4 GTexel/s
FP32 (TFLOPS)
74.45 TFLOPS
8.141 TFLOPS
FP64 (TFLOPS)
37.22 TFLOPS (1:2)
254.4 GFLOPS (1:32)
FP16 (TFLOPS)
1,191.2 TFLOPS (16:1)
65.13 TFLOPS (8:1)
AI/RT
RT Cores
—
40
Tensor Cores
592
320
-45.9%
Power
TDP
1000 W
70 W
TDP (W)
1,000
70
-93.0%
Suggested PSU
1400 W
250 W
Power Connectors
—
None
Architecture
Architecture
Blackwell
Turing
GPU Name
GB100
TU104
Generation
Server Blackwell
(Bxx)
Tesla Turing
(Txx)
Process Size
5 nm
12 nm
Transistors
104,000 million
13,600 million
Die Size
—
545 mm²
Foundry
TSMC
TSMC
Density
—
25.0M / mm²
API Support
DirectX
—
12 Ultimate (12_2)
OpenGL
—
4.6
Vulkan
—
1.4
OpenCL
3.0
3.0
CUDA
10.0
7.5
Shader Model
—
6.8
Physical
Slot Width
SXM Module
Single-slot
Length
—
168 mm
6.6 inches
Outputs
No outputs
No outputs
Bus Interface
PCIe 5.0 x16
PCIe 3.0 x16
Other
Production
Active
End-of-life
Predecessor
Server Hopper
Tesla Volta
Successor
Server Rubin
Server Ampere