NVIDIA H200 NVL vs NVIDIA GeForce GTX 1650 GDDR6
We compared two Desktop platform GPUs: 141GB VRAM H200 NVL and 4GB VRAM GeForce GTX 1650 GDDR6 to see which GPU has better performance in key specifications, benchmark tests, power consumption, etc.
Main Differences
NVIDIA H200 NVL 's Advantages
Released 4 years and 7 months late
Boost Clock has increased by 12% (1785MHz vs 1590MHz)
More VRAM (141GB vs 4GB)
Larger VRAM bandwidth (4.89TB/s vs 192.0GB/s)
16000 additional rendering cores
NVIDIA GeForce GTX 1650 GDDR6 's Advantages
Lower TDP (75W vs 600W)
Score
Benchmark
FP32 (float)
H200 NVL
+2017%
60.32 TFLOPS
GeForce GTX 1650 GDDR6
2.849 TFLOPS
Graphics Card
Nov 2024
Release Date
Apr 2020
Server Hopper
Generation
GeForce 16
Desktop
Type
Desktop
PCIe 5.0 x16
Bus Interface
PCIe 3.0 x16
Clock Speeds
1365 MHz
Base Clock
1410 MHz
1785 MHz
Boost Clock
1590 MHz
1593 MHz
Memory Clock
1500 MHz
Memory
141GB
Memory Size
4GB
HBM3e
Memory Type
GDDR6
6144bit
Memory Bus
128bit
4.89TB/s
Bandwidth
192.0GB/s
Render Config
-
-
-
132
SM Count
14
16896
Shading Units
896
528
TMUs
56
24
ROPs
32
528
Tensor Cores
-
-
-
-
256 KB (per SM)
L1 Cache
64 KB (per SM)
50 MB
L2 Cache
1024 KB
-
-
-
Theoretical Performance
42.84 GPixel/s
Pixel Rate
50.88 GPixel/s
942.5 GTexel/s
Texture Rate
89.04 GTexel/s
241.3 TFLOPS
FP16 (half)
5.699 TFLOPS
60.32 TFLOPS
FP32 (float)
2.849 TFLOPS
30.16 TFLOPS
FP64 (double)
89.04 GFLOPS
Board Design
600W
TDP
75W
1000 W
Suggested PSU
250 W
No outputs
Outputs
1x DVI
1x HDMI 2.0
1x DisplayPort 1.4a
8-pin EPS
Power Connectors
None
Graphics Processor
GH100
GPU Name
TU117
-
GPU Variant
TU117-300-A1
Hopper
Architecture
Turing
TSMC
Foundry
TSMC
5 nm
Process Size
12 nm
80 billion
Transistors
4.7 billion
814 mm²
Die Size
200 mm²
Graphics Features
N/A
DirectX
12 (12_1)
N/A
OpenGL
4.6
3.0
OpenCL
3.0
N/A
Vulkan
1.3
9.0
CUDA
7.5
N/A
Shader Model
6.6
