NVIDIA CMP 40HX vs NVIDIA H200 NVL
We compared two Desktop platform GPUs: 8GB VRAM CMP 40HX and 141GB VRAM H200 NVL to see which GPU has better performance in key specifications, benchmark tests, power consumption, etc.
Main Differences
NVIDIA CMP 40HX 's Advantages
Lower TDP (185W vs 600W)
NVIDIA H200 NVL 's Advantages
Released 3 years and 9 months late
Boost Clock has increased by 8% (1785MHz vs 1650MHz)
More VRAM (141GB vs 8GB)
Larger VRAM bandwidth (4.89TB/s vs 448.0GB/s)
14592 additional rendering cores
Score
Benchmark
FP32 (float)
CMP 40HX
7.603 TFLOPS
H200 NVL
+693%
60.32 TFLOPS
Graphics Card
Feb 2021
Release Date
Nov 2024
Mining GPUs
Generation
Server Hopper
Desktop
Type
Desktop
PCIe 1.0 x4
Bus Interface
PCIe 5.0 x16
Clock Speeds
1470 MHz
Base Clock
1365 MHz
1650 MHz
Boost Clock
1785 MHz
1750 MHz
Memory Clock
1593 MHz
Memory
8GB
Memory Size
141GB
GDDR6
Memory Type
HBM3e
256bit
Memory Bus
6144bit
448.0GB/s
Bandwidth
4.89TB/s
Render Config
-
-
-
36
SM Count
132
2304
Shading Units
16896
144
TMUs
528
64
ROPs
24
288
Tensor Cores
528
36
RT Cores
-
64 KB (per SM)
L1 Cache
256 KB (per SM)
4 MB
L2 Cache
50 MB
-
-
-
Theoretical Performance
105.6 GPixel/s
Pixel Rate
42.84 GPixel/s
237.6 GTexel/s
Texture Rate
942.5 GTexel/s
15.21 TFLOPS
FP16 (half)
241.3 TFLOPS
7.603 TFLOPS
FP32 (float)
60.32 TFLOPS
237.6 GFLOPS
FP64 (double)
30.16 TFLOPS
Board Design
185W
TDP
600W
450 W
Suggested PSU
1000 W
No outputs
Outputs
No outputs
1x 8-pin
Power Connectors
8-pin EPS
Graphics Processor
TU106
GPU Name
GH100
TU106-100-A1
GPU Variant
-
Turing
Architecture
Hopper
TSMC
Foundry
TSMC
12 nm
Process Size
5 nm
10.8 billion
Transistors
80 billion
445 mm²
Die Size
814 mm²
Graphics Features
12 Ultimate (12_2)
DirectX
N/A
4.6
OpenGL
N/A
3.0
OpenCL
3.0
1.3
Vulkan
N/A
7.5
CUDA
9.0
6.8
Shader Model
N/A
