Home GPU Comparison NVIDIA Tesla K20s vs NVIDIA H100 PCIe 96 GB

NVIDIA Tesla K20s vs NVIDIA H100 PCIe 96 GB

AI GPU We compared a Professional market GPU: 5GB VRAM Tesla K20s and a GPU: 96GB VRAM H100 PCIe 96 GB to see which GPU has better performance in key specifications, benchmark tests, power consumption, etc.

Main Differences

NVIDIA Tesla K20s 's Advantages
Lower TDP (225W vs 700W)
NVIDIA H100 PCIe 96 GB 's Advantages
Released 10 years and 1 months late
Boost Clock1837MHz
More VRAM (96GB vs 5GB)
Larger VRAM bandwidth (1681GB/s vs 208.0GB/s)
14400 additional rendering cores

Score

Benchmark

FP32 (float)
Tesla K20s
2.87 TFLOPS
H100 PCIe 96 GB +2063%
62.08 TFLOPS
VS

Graphics Card

Feb 2013
Release Date
Mar 2023
Tesla
Generation
Tesla Hopper
Professional
Type
AI GPU
PCIe 2.0 x16
Bus Interface
PCIe 5.0 x16

Clock Speeds

-
Base Clock
1665 MHz
-
Boost Clock
1837 MHz
1300 MHz
Memory Clock
1313 MHz

Memory

5GB
Memory Size
96GB
GDDR5
Memory Type
HBM3
320bit
Memory Bus
5120bit
208.0GB/s
Bandwidth
1681GB/s

Render Config

-
SM Count
132
-
Compute Units
-
2496
Shading Units
16896
208
TMUs
528
40
ROPs
24
-
Tensor Cores
528
-
RT Cores
-
16 KB (per SMX)
L1 Cache
256 KB (per SM)
1280 KB
L2 Cache
50 MB

Theoretical Performance

29.90 GPixel/s
Pixel Rate
44.09 GPixel/s
119.6 GTexel/s
Texture Rate
969.9 GTexel/s
-
FP16 (half)
248.3 TFLOPS
2.870 TFLOPS
FP32 (float)
62.08 TFLOPS
956.8 GFLOPS
FP64 (double)
31.04 TFLOPS

Graphics Processor

GK110
GPU Name
GH100
-
GPU Variant
-
Kepler
Architecture
Hopper
TSMC
Foundry
TSMC
28 nm
Process Size
4 nm
7.08 billion
Transistors
80 billion
561 mm²
Die Size
814 mm²

Board Design

225W
TDP
700W
550 W
Suggested PSU
1100 W
No outputs
Outputs
No outputs
-
Power Connectors
8-pin EPS

Graphics Features

12 (11_0)
DirectX
N/A
4.6
OpenGL
N/A
3.0
OpenCL
3.0
1.1
Vulkan
N/A
3.5
CUDA
9.0
5.1
Shader Model
N/A

Related GPU Comparisons

© 2024 - TopCPU.net   Contact Us Privacy Policy