NVIDIA L40S vs NVIDIA H100 PCIe 80 GB

Comparative analysis of NVIDIA L40S and NVIDIA H100 PCIe 80 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: Geekbench - OpenCL.

NVIDIA L40S

NVIDIA H100 PCIe 80 GB

Differences

Reasons to consider the NVIDIA L40S

Videocard is newer: launch date 6 month(s) later
Around 1% higher core clock speed: 1110 MHz vs 1095 MHz
Around 44% higher boost clock speed: 2520 MHz vs 1755 MHz
Around 25% higher pipelines: 18176 vs 14592
Around 17% lower typical power consumption: 300 Watt vs 350 Watt
Around 41% higher memory clock speed: 2250 MHz, 18 Gbps effective vs 1593 MHz, 3.2 Gbps effective
Around 25% better performance in Geekbench - OpenCL: 352097 vs 281732

Specifications (specs)
Launch date	13 Oct 2022 vs 22 Mar 2022
Core clock speed	1110 MHz vs 1095 MHz
Boost clock speed	2520 MHz vs 1755 MHz
Pipelines	18176 vs 14592
Thermal Design Power (TDP)	300 Watt vs 350 Watt
Memory clock speed	2250 MHz, 18 Gbps effective vs 1593 MHz, 3.2 Gbps effective
Benchmarks
Geekbench - OpenCL	352097 vs 281732

Reasons to consider the NVIDIA H100 PCIe 80 GB

559.3x more texture fill rate: 800.3 GTexel/s vs 1,431 GTexel/s
A newer manufacturing process allows for a more powerful, yet cooler running videocard: 4 nm vs 5 nm
Around 67% higher maximum memory size: 80 GB vs 48 GB

Texture fill rate	800.3 GTexel/s vs 1,431 GTexel/s
Manufacturing process technology	4 nm vs 5 nm
Maximum memory size	80 GB vs 48 GB

Compare benchmarks

GPU 1: NVIDIA L40S
GPU 2: NVIDIA H100 PCIe 80 GB

Geekbench - OpenCL

GPU 1

GPU 2

352097

281732

Name	NVIDIA L40S	NVIDIA H100 PCIe 80 GB
Geekbench - OpenCL	352097	281732

Compare specifications (specs)

	NVIDIA L40S	NVIDIA H100 PCIe 80 GB
Essentials
Architecture	Ada Lovelace	Hopper
Code name	AD102	GH100
Launch date	13 Oct 2022	22 Mar 2022
Place in performance rating	1	5
Technical info
Boost clock speed	2520 MHz	1755 MHz
Core clock speed	1110 MHz	1095 MHz
Manufacturing process technology	5 nm	4 nm
Pipelines	18176	14592
Pixel fill rate	483.8 GPixel/s	42.12 GPixel/s
Texture fill rate	1,431 GTexel/s	800.3 GTexel/s
Thermal Design Power (TDP)	300 Watt	350 Watt
Transistor count	76300 million	80000 million
Video outputs and ports
Display Connectors	1x HDMI 2.1, 3x DisplayPort 1.4a	No outputs
Compatibility, dimensions and requirements
Form factor	Dual-slot	Dual-slot
Interface	PCIe 4.0 x16	PCIe 5.0 x16
Length	267 mm, 10.5 inches	268 mm, 10.6 inches
Recommended system power (PSU)	700 Watt	750 Watt
Supplementary power connectors	1x 16-pin	1x 16-pin
Width	111 mm, 4.4 inches	111 mm, 4.4 inches
API support
DirectX	12 Ultimate (12_2)
OpenCL	3.0	3.0
OpenGL	4.6
Shader Model	6.7
Vulkan
Memory
Maximum RAM amount	48 GB	80 GB
Memory bandwidth	864.0 GB/s	2,039 GB/s
Memory bus width	384 bit	5120 bit
Memory clock speed	2250 MHz, 18 Gbps effective	1593 MHz, 3.2 Gbps effective
Memory type	GDDR6	HBM2e