NVIDIA Tesla K80 vs NVIDIA A100 SXM4 40 GB

Comparative analysis of NVIDIA Tesla K80 and NVIDIA A100 SXM4 40 GB videocards for all known characteristics in the following categories: Essentials, Technical info, Video outputs and ports, Compatibility, dimensions and requirements, API support, Memory. Benchmark videocards performance analysis: PassMark - G2D Mark, PassMark - G3D Mark, Geekbench - OpenCL, GFXBench 4.0 - Car Chase Offscreen (Frames), GFXBench 4.0 - Car Chase Offscreen (Fps), GFXBench 4.0 - Manhattan (Frames), GFXBench 4.0 - Manhattan (Fps), GFXBench 4.0 - T-Rex (Frames), GFXBench 4.0 - T-Rex (Fps).

NVIDIA Tesla K80

NVIDIA A100 SXM4 40 GB

Differences

Reasons to consider the NVIDIA Tesla K80

Around 33% lower typical power consumption: 300 Watt vs 400 Watt
Around 3% higher memory clock speed: 1253 MHz, 5 Gbps effective vs 1215 MHz (2.4 Gbps effective)

Thermal Design Power (TDP)	300 Watt vs 400 Watt
Memory clock speed	1253 MHz, 5 Gbps effective vs 1215 MHz (2.4 Gbps effective)

Reasons to consider the NVIDIA A100 SXM4 40 GB

Videocard is newer: launch date 5 year(s) 5 month(s) later
Around 95% higher core clock speed: 1095 MHz vs 562 MHz
Around 71% higher boost clock speed: 1410 MHz vs 824 MHz
3.6x more texture fill rate: 609.1 GTexel/s vs 171.4 GTexel/s
2.8x more pipelines: 6912 vs 2496
A newer manufacturing process allows for a more powerful, yet cooler running videocard: 7 nm vs 28 nm
3.3x more maximum memory size: 40 GB vs 12 GB

Launch date	14 May 2020 vs 17 Nov 2014
Core clock speed	1095 MHz vs 562 MHz
Boost clock speed	1410 MHz vs 824 MHz
Texture fill rate	609.1 GTexel/s vs 171.4 GTexel/s
Pipelines	6912 vs 2496
Manufacturing process technology	7 nm vs 28 nm
Maximum memory size	40 GB vs 12 GB

Compare benchmarks

GPU 1: NVIDIA Tesla K80
GPU 2: NVIDIA A100 SXM4 40 GB

Name	NVIDIA Tesla K80	NVIDIA A100 SXM4 40 GB
PassMark - G2D Mark	280
PassMark - G3D Mark	4680
Geekbench - OpenCL		200738
GFXBench 4.0 - Car Chase Offscreen (Frames)		21006
GFXBench 4.0 - Car Chase Offscreen (Fps)		21006
GFXBench 4.0 - Manhattan (Frames)		27823
GFXBench 4.0 - Manhattan (Fps)		27823
GFXBench 4.0 - T-Rex (Frames)		51880
GFXBench 4.0 - T-Rex (Fps)		51880

Compare specifications (specs)

	NVIDIA Tesla K80	NVIDIA A100 SXM4 40 GB
Essentials
Architecture	Kepler 2.0	Ampere
Code name	GK210	GA100
Launch date	17 Nov 2014	14 May 2020
Place in performance rating	374	7
Technical info
Boost clock speed	824 MHz	1410 MHz
Core clock speed	562 MHz	1095 MHz
Manufacturing process technology	28 nm	7 nm
Peak Double Precision (FP64) Performance	1,371 GFLOPS (1:3)	9.746 TFLOPS (1:2)
Peak Single Precision (FP32) Performance	4.113 TFLOPS	19.49 TFLOPS
Pipelines	2496	6912
Pixel fill rate	42.85 GPixel/s	225.6 GPixel/s
Texture fill rate	171.4 GTexel/s	609.1 GTexel/s
Thermal Design Power (TDP)	300 Watt	400 Watt
Transistor count	7100 million	54200 million
Peak Half Precision (FP16) Performance		77.97 TFLOPS (4:1)
Video outputs and ports
Display Connectors	No outputs	No outputs
Compatibility, dimensions and requirements
Form factor	Dual-slot	IGP
Interface	PCIe 3.0 x16	PCIe 4.0 x16
Length	267 mm, 10.5 inches
Recommended system power (PSU)	700 Watt	800 Watt
Supplementary power connectors	1x 8-pin	None
API support
DirectX	12 (11_1)
OpenCL	3.0	3.0
OpenGL	4.6
Shader Model	6.5 (5.1)
Vulkan
Memory
Maximum RAM amount	12 GB	40 GB
Memory bandwidth	240.6 GB/s	1555 GB/s
Memory bus width	384 bit	5120 bit
Memory clock speed	1253 MHz, 5 Gbps effective	1215 MHz (2.4 Gbps effective)
Memory type	GDDR5	HBM2e
High bandwidth memory (HBM)