Nvidia GeForce RTX 4090 vs Nvidia H100 PCIe 80GB

Nvidia GeForce RTX 4090 on Amazon USA
Nvidia H100 PCIe 80GB on Amazon USA

Comparison of the technical characteristics between the graphics cards, with Nvidia GeForce RTX 4090 on one side and Nvidia H100 PCIe 80GB on the other side, also their respective performances with the benchmarks. The first is dedicated to the desktop sector, it has 16834 shading units, a maximum frequency of 2.5 GHz, its lithography is 5 nm. The second is used on the desktop segment, it includes 14592 shading units, a maximum frequency of 1.8 GHz, its lithography is 4 nm. The following table also compares the boost clock, the number of shading units (if indicated), of execution units, the amount of cache memory, the maximum memory capacity, the memory bus width, the release date, the number of PCIe lanes, the values ​​obtained in various benchmarks.

Specification comparison:

Graphics card Nvidia GeForce RTX 4090 Nvidia H100 PCIe 80GB
Market (main) Desktop Desktop
Release date Q3 2022 Q1 2022
Model number AD102-300-A1 H100 PCIe 80GB
GPU name AD102 GH100
Architecture Ada Lovelace Hopper
Generation GeForce 40 Tesla Hopper Hxx
Lithography 5 nm 4 nm
Transistors 76,300,000,000 80,000,000,000
Bus interface PCIe 4.0 x16 PCIe 5.0 x16
GPU base clock 2,235 MHz 1,095 MHz
GPU boost clock 2,520 MHz 1,755 MHz
Memory frequency 1,325 MHz 1,593 MHz
Effective memory speed 21 Gbps 3.2 Gbps
Memory size 24 GB 80 GB
Memory type GDDR6X HBM2e
Memory bus 384 bit 5120 bit
Memory bandwidth 1,008.0 GB/s 2,039.0 GB/s
TDP 450 W 350 W
Suggested PSU850W ATX Power Supply 850W ATX Power Supply
Multicard technology - NVLink
Max display resolution

3840 × 2160@240Hz (DP),
7680 x 4320@60Hz (DP),
7680 x 4320@120Hz (2x DP),
7680 x 4320@60Hz (HDMI)

Video decoding H.264, AV1, HEVC -
Video encoding H.264, AV1, HEVC -
Cores (compute units, SM, SMX) 128 114
Shading units (cuda cores) 16,834 14,592
TMUs 512 456
ROPs 176 24
Tensor cores 512 456
Cache memory 72 MB 50 MB
Pixel fillrate 443.5 GP/s 42.1 GP/s
Texture fillrate 1,290.2 GT/s 800.3 GT/s
Performance FP16 (half) 82.6 TFLOPS 204.9 TFLOPS
Performance FP32 (float) 82.6 TFLOPS 51.2 TFLOPS
Performance FP64 (double) 1.3 TFLOPS 25.6 TFLOPS

Price: For technical reasons, we cannot currently display a price less than 24 hours, or a real-time price. This is why we prefer for the moment not to show a price. You should refer to the respective online stores for the latest price, as well as availability.

We can better compare what are the technical differences between the two graphics cards.

Performance comparison with the benchmarks:

Performance comparison between the two processors, for this we consider the results generated on benchmark software such as Geekbench.

FP32 Performance in GFLOPS
Nvidia GeForce RTX 4090 82,575
Nvidia H100 PCIe 80GB 51,218

The difference is 61%.

Single-precision floating-point format, also called FP32, is a computer number format that typically occupies 32 bits in PC memory. This represents a wide dynamic range of numeric values that employs a floating point. FP32 power helps assess the video card's capabilities for generating rendering workloads. Video games use INT32 and FP32 cores to manage calculations with integers and numbers with decimals in order to produce maps, bone matrices on which the vertex shaders will calculate the vertices. Values in single precision format are mainly used in video games and 3D software.

See also:

Nvidia GeForce RTX 4090 DNvidia GeForce RTX 4090 Founders EditionNvidia GeForce RTX 4090 Max-QNvidia GeForce RTX 4090 MobileNvidia GeForce RTX 4090 Ti


Nvidia GeForce RTX 4090 AMD EquivalentNvidia H100 PCIe 80GB AMD Equivalent

GadgetVersus.com © 2024 - About us - XML sitemap - Privacy policy - Advertising policy - Contact


When you click on links to various merchants on this site and make a purchase, this can result in this site earning a commission. Affiliate programs and affiliations include, but are not limited to, the eBay Partner Network.

As an Amazon Associate I earn from qualifying purchases.

This page includes affiliate links for which the administrator of GadgetVersus may earn a commission at no extra cost to you should you make a purchase. These links are indicated using the hashtag #ad.


We do not assume any responsibility for the data displayed on our website. Please use at your own risk. Some or all of this data may be out of date or incomplete, please refer to the technical page on the respective manufacturer's website to find the latest up-to-date information regarding the specifics of these products.