NVIDIA H200 NVL vs NVIDIA GeForce RTX 4070 GDDR6
We compared two Desktop platform GPUs: 141GB VRAM H200 NVL and 12GB VRAM GeForce RTX 4070 GDDR6 to see which GPU has better performance in key specifications, benchmark tests, power consumption, etc.
Main Differences
NVIDIA H200 NVL 's Advantages
More VRAM (141GB vs 12GB)
Larger VRAM bandwidth (4.89TB/s vs 480.0GB/s)
11008 additional rendering cores
NVIDIA GeForce RTX 4070 GDDR6 's Advantages
Boost Clock has increased by 39% (2475MHz vs 1785MHz)
Lower TDP (200W vs 600W)
Score
Benchmark
FP32 (float)
H200 NVL
+106%
60.32 TFLOPS
GeForce RTX 4070 GDDR6
29.15 TFLOPS
Graphics Card
Nov 2024
Release Date
Aug 2024
Server Hopper
Generation
GeForce 40
Desktop
Type
Desktop
PCIe 5.0 x16
Bus Interface
PCIe 4.0 x16
Clock Speeds
1365 MHz
Base Clock
1920 MHz
1785 MHz
Boost Clock
2475 MHz
1593 MHz
Memory Clock
2500 MHz
Memory
141GB
Memory Size
12GB
HBM3e
Memory Type
GDDR6
6144bit
Memory Bus
192bit
4.89TB/s
Bandwidth
480.0GB/s
Render Config
-
-
-
132
SM Count
46
16896
Shading Units
5888
528
TMUs
184
24
ROPs
64
528
Tensor Cores
184
-
RT Cores
46
256 KB (per SM)
L1 Cache
128 KB (per SM)
50 MB
L2 Cache
36 MB
-
-
-
Theoretical Performance
42.84 GPixel/s
Pixel Rate
158.4 GPixel/s
942.5 GTexel/s
Texture Rate
455.4 GTexel/s
241.3 TFLOPS
FP16 (half)
29.15 TFLOPS
60.32 TFLOPS
FP32 (float)
29.15 TFLOPS
30.16 TFLOPS
FP64 (double)
455.4 GFLOPS
Board Design
600W
TDP
200W
1000 W
Suggested PSU
550 W
No outputs
Outputs
1x HDMI 2.1
3x DisplayPort 1.4a
8-pin EPS
Power Connectors
1x 16-pin
Graphics Processor
GH100
GPU Name
AD104
-
GPU Variant
AD104-2XX-A1
Hopper
Architecture
Ada Lovelace
TSMC
Foundry
TSMC
5 nm
Process Size
5 nm
80 billion
Transistors
35.8 billion
814 mm²
Die Size
294 mm²
Graphics Features
N/A
DirectX
12 Ultimate (12_2)
N/A
OpenGL
4.6
3.0
OpenCL
3.0
N/A
Vulkan
1.3
9.0
CUDA
8.9
N/A
Shader Model
6.8
