GeekRadar / The comparison desk
COMPARE.
GPU
NVIDIA

NVIDIA B100

Graphics processor

Video memory
192 GB*
Memory bus
4,096 bit*
Rated memory bandwidth
4,100 GB/s*

No matched benchmark results are available for this entry yet.

Processing & platform

Architecture
Blackwell*
Process node
5 nm*
GPU cores
10,240*
Graphics interface
PCIe 6.0 x16*
Reported chip die area
814 mm²*
Transistors
104,000 million*
GPU L2 cache
64 MB*
Graphics processor
GB102*
Texture mapping units
320*
Raster output units
24*
Streaming multiprocessors
80*
Tensor cores
320*
Chip foundry
TSMC*
Card or module format
SXM Module*
Declared GPU TDP
1,000 W*
Suggested system PSU
1,400 W*
OpenCL
3.0*
CUDA compute capability
10.1*

Memory & storage

Memory options
96*
Video memory
192 GB*
Video memory type
HBM3e*
Memory bus
4,096 bit*
Rated memory bandwidth
4,100 GB/s*

Power specifications

Thermal design power
1,000 W*

System configuration

Detailed specifications

Specifications 1 · 27 fields
proc transistors
104,000 million*
cores
16,896*
tensor cores
528*
sm count
132*
base clock
700*
boost clock
975*
memory clock
2,000*
ram
96*
name
NVIDIA B100*
ram type
HBM3e*
bus width
4,096*
bus interface
PCIe 5.0 x16*
ram bandwidth
4.10 TB/s*
fp16 performance
131.8*
fp32 performance
32.95*
fp64 performance
16.47*
tdp
1,000* W
suggested psu
1,400* W
slot width
SXM Module*
api cuda
10.1*
is retail board
No*
gpi value
42.44*
vendor
nvd*
manufacturer
NVIDIA*
arch name
Blackwell*
proc foundry
TSMC*
proc process size
5* nm
Specifications 2 · 33 fields
CUDA compute capability
10.1*
OpenCL
3.0*
Graphics interface
PCIe 6.0 x16*
Memory bus
4,096* bit
Reported shader processors
10,240*
Reported theoretical FP16 peak
79.87* TFLOPS
Scope and conditions
Arithmetic density
Not stated*
Reported theoretical FP32 peak
19.97* TFLOPS
Scope and conditions
Arithmetic density
Not stated*
Reported theoretical FP64 peak
9.984* TFLOPS
Scope and conditions
Arithmetic density
Not stated*
Declared L1 cache configuration
256 KB (per SM)*
Declared GPU L2 cache
64 MB*
Reported display outputs
No outputs*
Theoretical pixel fill rate
23.40 GPixel/s*
Reported chip die area
814 mm²*
Chip foundry
TSMC*
Process node declaration
5 nm*
Transistors
104,000 million*
Graphics processor
GB102*
Reported video memory
192* GB
Rated memory bandwidth
4.10 TB/s*
Video memory type
HBM3e*
Raster output units
24*
Reported shading units
10,240*
Card or module format
SXM Module*
Streaming multiprocessors
80*
Suggested system PSU
1400 W*
2:4 tensor sparsity
No*
Declared GPU TDP
1000 W*
Tensor cores
320*
Theoretical texture fill rate
312.0 GTexel/s*
Texture mapping units
320*
GPU cores
10,240*
Video memory
192* GB
Video memory type
HBM3e*

* Provisional data