NVIDIA H200 SXM 141 GB vs NVIDIA GeForce 8800 GT
We compared two Desktop platform GPUs: 141GB VRAM H200 SXM 141 GB and 512MB VRAM GeForce 8800 GT to see which GPU has better performance in key specifications, benchmark tests, power consumption, etc.
Main Differences
NVIDIA H200 SXM 141 GB 's Advantages
Released 17 years and 1 months late
Boost Clock1980MHz
More VRAM (141GB vs 512GB)
Larger VRAM bandwidth (4.89TB/s vs 57.60GB/s)
16784 additional rendering cores
NVIDIA GeForce 8800 GT 's Advantages
Lower TDP (125W vs 700W)
Score
Benchmark
FP32 (float)
H200 SXM 141 GB
+19813%
66.91 TFLOPS
GeForce 8800 GT
0.336 TFLOPS
Graphics Card
Nov 2024
Release Date
Oct 2007
Server Hopper
Generation
GeForce 8
Desktop
Type
Desktop
PCIe 5.0 x16
Bus Interface
PCIe 2.0 x16
Clock Speeds
1500 MHz
Base Clock
-
1980 MHz
Boost Clock
-
1593 MHz
Memory Clock
900 MHz
Memory
141GB
Memory Size
512MB
HBM3e
Memory Type
GDDR3
6144bit
Memory Bus
256bit
4.89TB/s
Bandwidth
57.60GB/s
Render Config
-
-
-
132
SM Count
14
16896
Shading Units
112
528
TMUs
56
24
ROPs
16
528
Tensor Cores
-
-
-
-
256 KB (per SM)
L1 Cache
-
50 MB
L2 Cache
64 KB
-
-
-
Theoretical Performance
47.52 GPixel/s
Pixel Rate
9.600 GPixel/s
1045 GTexel/s
Texture Rate
33.60 GTexel/s
267.6 TFLOPS
FP16 (half)
-
66.91 TFLOPS
FP32 (float)
336.0 GFLOPS
33.45 TFLOPS
FP64 (double)
-
Board Design
700W
TDP
125W
1100 W
Suggested PSU
300 W
No outputs
Outputs
2x DVI
1x S-Video
8-pin EPS
Power Connectors
1x 6-pin
Graphics Processor
GH100
GPU Name
G92
-
GPU Variant
G92-270-A2
Hopper
Architecture
Tesla
TSMC
Foundry
TSMC
5 nm
Process Size
65 nm
80 billion
Transistors
0.754 billion
814 mm²
Die Size
324 mm²
Graphics Features
N/A
DirectX
11.1 (10_0)
N/A
OpenGL
3.3
3.0
OpenCL
1.1
N/A
Vulkan
N/A
9.0
CUDA
1.1
N/A
Shader Model
4.0
