GPU Comparison

NVIDIA

GEFORCE

NVIDIA B200

CORE STATE GB100

VRAM 90 GB

CLOCK SPEED 1965 MHz

TDP 1000 W

BUS WIDTH 4096 bit

ARCHITECTURE Blackwell

nm

PROCESS 5 nm

LAUNCH DATE —

VS

NVIDIA

GEFORCE

Quadro P4000

CORE STATE GP104

VRAM 8 GB

CLOCK SPEED 1480 MHz

TDP 105 W

BUS WIDTH 256 bit

ARCHITECTURE Pascal

nm

PROCESS 16 nm

LAUNCH DATE 2017

PERFORMANCE BENCHMARKS

geekbench_opencl

345,482

35,713

3dmark_3dmark_steel_nomad_dx12

N/A

1,115

geekbench_vulkan

N/A

41,786

passmark_directx_10

N/A

66

passmark_directx_11

N/A

86

passmark_directx_12

N/A

40

passmark_directx_9

N/A

181

passmark_g2d

N/A

786

passmark_g3d

N/A

11,466

passmark_gpu_compute

N/A

4,913

DETAILED SPECIFICATIONS

SPECIFICATION

B200

Quadro P4000

Core Specs

Shading Units

18,944

1,792 -90.5%

Shaders

18,944

1,792 -90.5%

TMUs

592

112 -81.1%

ROPs

24

64 +166.7%

SM Count

148

14 -90.5%

Clocks

Base Clock

700 MHz

1202 MHz

Boost Clock

1965 MHz

1480 MHz

Memory Clock

2000 MHz 8 Gbps effective

1901 MHz 7.6 Gbps effective

Memory

Memory Size

90 GB

8 GB

VRAM (MB)

92,160

8,192 -91.1%

Memory Type

HBM3e

GDDR5

Memory Bus

4096 bit

256 bit

Bandwidth

4.10 TB/s

243.3 GB/s

Cache

L1 Cache

256 KB (per SM)

48 KB (per SM)

L2 Cache

50 MB

2 MB

Performance

Pixel Rate

47.16 GPixel/s

94.72 GPixel/s

Texture Rate

1,163.3 GTexel/s

165.8 GTexel/s

FP32 (TFLOPS)

74.45 TFLOPS

5.304 TFLOPS

FP64 (TFLOPS)

37.22 TFLOPS (1:2)

165.8 GFLOPS (1:32)

FP16 (TFLOPS)

1,191.2 TFLOPS (16:1)

82.88 GFLOPS (1:64)

AI/RT

Tensor Cores

592

—

Power

TDP

1000 W

105 W

TDP (W)

1,000

105 -89.5%

Suggested PSU

1400 W

300 W

Power Connectors

—

1x 6-pin

Architecture

Architecture

Blackwell

Pascal

GPU Name

GB100

GP104

Generation

Server Blackwell (Bxx)

Quadro Pascal (Px000)

Process Size

5 nm

16 nm

Transistors

104,000 million

7,200 million

Die Size

—

314 mm²

Foundry

TSMC

TSMC

Density

—

22.9M / mm²

API Support

DirectX

—

12 (12_1)

OpenGL

—

4.6

Vulkan

—

1.4

OpenCL

3.0

3.0

CUDA

10.0

6.1

Shader Model

—

6.8

Physical

Slot Width

SXM Module

Single-slot

Length

—

241 mm 9.5 inches

Height

—

111 mm 4.4 inches

Outputs

No outputs

4x DisplayPort 1.4a

Bus Interface

PCIe 5.0 x16

PCIe 3.0 x16

Other

Launch Price

—

815 USD

Production

Active

End-of-life

Predecessor

Server Hopper

Quadro Maxwell

Successor

Server Rubin

Quadro Volta

View B200 Details View Quadro P4000 Details