GPU Comparison

NVIDIA

GEFORCE

NVIDIA GeForce RTX 5070

CORE STATE GB205

VRAM 12 GB

CLOCK SPEED 2512 MHz

TDP 250 W

BUS WIDTH 192 bit

ARCHITECTURE Blackwell 2.0

nm

PROCESS 5 nm

LAUNCH DATE 2025

VS

NVIDIA

GEFORCE

H200 NVL

CORE STATE GH100

VRAM 141 GB

CLOCK SPEED 1785 MHz

TDP 600 W

BUS WIDTH 6144 bit

ARCHITECTURE Hopper

nm

PROCESS 5 nm

LAUNCH DATE 2024

PERFORMANCE BENCHMARKS

3dmark_3dmark_steel_nomad_dx12

5,077

N/A

geekbench_opencl

185,269

305,608

geekbench_vulkan

179,413

N/A

passmark_directx_10

180

N/A

passmark_directx_11

277

N/A

passmark_directx_12

108

N/A

passmark_directx_9

320

N/A

passmark_g2d

1,305

N/A

passmark_g3d

29,137

N/A

passmark_gpu_compute

15,787

N/A

DETAILED SPECIFICATIONS

SPECIFICATION

RTX 5070

H200 NVL

Core Specs

Shading Units

6,144

16,896 +175.0%

Shaders

6,144

16,896 +175.0%

TMUs

192

528 +175.0%

ROPs

80

24 -70.0%

SM Count

48

132 +175.0%

Clocks

Base Clock

2325 MHz

1365 MHz

Boost Clock

2512 MHz

1785 MHz

Memory Clock

1750 MHz 28 Gbps effective

1593 MHz 6.4 Gbps effective

Memory

Memory Size

12 GB

141 GB

VRAM (MB)

12,288

144,384 +1075.0%

Memory Type

GDDR7

HBM3e

Memory Bus

192 bit

6144 bit

Bandwidth

672.0 GB/s

4.89 TB/s

Cache

L1 Cache

128 KB (per SM)

256 KB (per SM)

L2 Cache

48 MB

50 MB

Performance

Pixel Rate

201.0 GPixel/s

42.84 GPixel/s

Texture Rate

482.3 GTexel/s

942.5 GTexel/s

FP32 (TFLOPS)

30.87 TFLOPS

60.32 TFLOPS

FP64 (TFLOPS)

482.3 GFLOPS (1:64)

30.16 TFLOPS (1:2)

FP16 (TFLOPS)

30.87 TFLOPS (1:1)

241.3 TFLOPS (4:1)

AI/RT

RT Cores

48

—

Tensor Cores

192

528 +175.0%

Power

TDP

250 W

600 W

TDP (W)

250

600 +140.0%

Suggested PSU

600 W

1000 W

Power Connectors

1x 16-pin

8-pin EPS

Architecture

Architecture

Blackwell 2.0

Hopper

GPU Name

GB205

GH100

Generation

GeForce 50

Server Hopper (Hxx)

Process Size

5 nm

5 nm

Transistors

31,100 million

80,000 million

Die Size

263 mm²

814 mm²

Foundry

TSMC

TSMC

Density

118.3M / mm²

98.3M / mm²

API Support

DirectX

12 Ultimate (12_2)

—

OpenGL

4.6

—

Vulkan

1.4

—

OpenCL

3.0

3.0

CUDA

12.0

9.0

Shader Model

6.8

—

Physical

Slot Width

Dual-slot

Dual-slot

Length

245 mm 9.6 inches

267 mm 10.5 inches

Height

115 mm 4.5 inches

111 mm 4.4 inches

Outputs

1x HDMI 2.1b3x DisplayPort 2.1b

No outputs

Bus Interface

PCIe 5.0 x16

PCIe 5.0 x16

Other

Launch Price

549 USD

—

Production

Active

Active

Predecessor

GeForce 40

Server Ada

Successor

—

Server Blackwell

View GeForce RTX 5070 Details View H200 NVL Details