NVIDIA GeForce RTX 3090 in 19 Benchmarks. Specs

It is time to review NVIDIA GeForce RTX 3090 benchmark performance. The base clock speed is 1400 MHz, whilst its speed in turbo boost and supplementary overclocking (maximum) is 1700 MHz.

NVIDIA GeForce RTX 3090 process technology is 8 nm that is a good result of the release date.

NVIDIA manufacture has equipped graphic chips with 936.2 GB/s bandwidth and 1400 MHz clock speed. NVIDIA GeForce RTX 3090 maximum memory is 24 GB.

We recommend you learn all technical specifications in separate sections for a common understanding and high-speed of NVIDIA GeForce RTX 3090.

General Information

Here is the release date, code name, core architecture, launch and current price. It is the general information to start learning NVIDIA GeForce RTX 3090.

Place in performance rating: 3
Value for money (0-100): 10.70
Architecture: Ampere
Code name: Ampere GA102
Release date: 24 September 2020 (1 year ago)
Launch price (MSRP): $1,499
Price now: $2574 (1.7x MSRP)
Market segment: Desktop

NVIDIA GeForce RTX 3090 - Technical Specifications

Here are key features that determine NVIDIA GeForce RTX 3090 performance in benchmarks and other apps. The higher the figures are, the better it is. Please pay attention to all the characteristics provided. For example, the clock frequency directly affects the high-speed, and processing data and performing calculations (e.g, for mining) depend on the number of transistors. Also, check the TDP figure, because modern graphics cards may require a powerful PSU.

Core clock speed: 1400 MHz
Boost Clock: 1700 MHz
Manufacturing process technology: 8 nm
Power consumption (TDP): 350 Watt
Texture fill rate: 556.0
CUDA cores per GPU: 10496
Number of transistors: 28,300 million

NVIDIA GeForce RTX 3090 - Dimensions and Compatibility

Nowadays, it is very important to know the dimensions - length, width and height of a desktop graphics card - and understand which power connector is included in the package. Also, please pay attention to the PCIe interface.

Interface: PCIe 4.0 x16
Length: 313 mm
Supplementary power connectors: 1x 12-pin

NVIDIA GeForce RTX 3090 - GPU RAM Frequency and Speed

Graphics Card Ram is a key element in computing and data storage - type, size, clock frequency and bandwidth - that directly affect the speed of work in gaming and working apps, which are currently quite demanding on graphics card resources. Note please the memory standard. The higher it is, the better it is. Let's learn more about NVIDIA GeForce RTX 3090 video memory specifications.

Memory type: GDDR6X
Maximum RAM amount: 24 GB
Memory bus width: 384 Bit
Memory clock speed: 19500 MHz
Memory bandwidth: 936.2 GB/s
Shared memory: -

NVIDIA GeForce RTX 3090 - Port and Display Support

In most cases, today all the problems connector compatibility can be solved with adapters or backwards-compatible cables such as HDMI / DVI. But, for general information, you learn below which ports are supported by NVIDIA GeForce RTX 3090.

Display Connectors: 1x HDMI, 3x DisplayPort
HDMI: +

NVIDIA GeForce RTX 3090 - Technologies

Healthy competition drives progress, so each manufacturer tries to surprise its consumer with special technologies available in this graphics card. This can be useful in graphics or gaming apps.

CUDA: 8.5

NVIDIA GeForce RTX 3090 - API support

Here are general standards for DirectX and other unified complexes for the development of 3D images supported by NVIDIA GeForce RTX 3090 .

DirectX: 12 Ultimate (12_2)
OpenGL: 4.6
Vulkan: 1.2
Shader Model: 6.5
OpenCL: 2.0

Overall NVIDIA GeForce RTX 3090 Gaming Performance

Find out to what extent NVIDIA GeForce RTX 3090 is suitable for gaming! We have compiled a database of popular games and their system requirements* that allows you to quickly and easily estimate its gaming performance at different screen resolutions. Graphics card gaming test is based on FPS. * There is a little margin error in the estimation of data.


low
1280x720
med.
1920x1080
high
1920x1080
ultra
1920x1080
QHD
2560x1440
4K
3840x2160
Horizon Zero Dawn (2020)
low
1280x720
med.
1920x1080
high
1920x1080
ultra
1920x1080
QHD
2560x1440
4K
3840x2160
Death Stranding (2020)
low
1280x720
med.
1920x1080
high
1920x1080
ultra
1920x1080
QHD
2560x1440
4K
3840x2160
F1 2020 (2020)
low
1280x720
med.
1920x1080
high
1920x1080
ultra
1920x1080
QHD
2560x1440
4K
3840x2160
Gears Tactics (2020)
low
1280x720
med.
1920x1080
high
1920x1080
ultra
1920x1080
QHD
2560x1440
4K
3840x2160
Doom Eternal (2020)
low
1280x720
med.
1920x1080
high
1920x1080
ultra
1920x1080
QHD
2560x1440
4K
3840x2160
Legend
5 Stutter – How graphics cards operate in the game is being examined for the time being. As per interpolation of the similar graphics cards showing low frame rates, the game is possible to lag.
May Stutter – How graphics cards operate in the game is being examined for the time being. As per interpolation of the similar graphics cards showing low frame rates, the game is possible to lag.
30 Fluent – In considering all graphics benchmarks, the game speed is likely to be 25 fps or more.
40 Fluent – In considering all graphics benchmarks, the game speed is likely to be 35 fps or more.
60 Fluent – In considering all graphics benchmarks, the game speed is likely to be 58 fps or more.
May Run Fluently – How graphics cards operate in the game is being examined for the time being. As per interpolation of the similar graphics cards showing low frame rates, the game is possible to lag.
? Uncertain – The graphics card testing has revealed a slower card might ensure a high and stable frame rate.
Uncertain – How graphics cards operate in the game is being examined for the time being. The interpolation in the same category is not possible.
The value depicts the average frame rate. Hovering over the value, you can get individual results.

NVIDIA GeForce RTX 3090 in Benchmarks

Let's sum up the review by comparing the benchmarks results of NVIDIA GeForce RTX 3090 5 competitive models similar in technical specifications. The first block is the overall result for all benchmarks, then all the others, which this particular model was tested on. The higher the benchmark score, the faster a video card performs in graphics apps and modern games. If you have any additional questions, please leave them in the comments, we will definitely help!

Overall benchmark performance

3DMark Ice Storm GPU

3DMark Ice Storm is the only fully cross-platform benchmark that runs on all devices. It's great for comparing the 3D performance of smartphones, tablets, ultraportable laptops, and old or low-end PCs.

3DMark Cloud Gate GPU

The 3DMark Cloud Gate GPU test is designed for mid-range laptops and home PCs with integrated video graphics. The test uses the DirectX 11 engine, which is limited by Direct3D 10 capabilities. Accordingly, Cloud Gate is best suited for testing modern entry-level systems and legacy DX10-enabled GPUs.

3DMark Fire Strike Score

Fire Strike, Fire Strike Extreme, Fire Strike Ultra. Fire Strike includes two graphics tests, a physics test and a combined test that stresses the CPU and GPU.

3DMark Fire Strike Graphics

3DMark Fire Strike for Graphics Cards helps you determine performance in a state-of-the-art synthetic test that consists of two levels - physical and combined testing.

3DMark 11 Performance GPU

The result of the 3DMark 11 benchmark evaluation supports tessellation, compute shaders and graphics card multithreading will show which graphics card model is better and faster in games. The higher the score of the video card in the 3DMark 11 benchmark, the better.

3DMark Vantage Performance

3DMark Vantage Performance reveals the real performance of video cards tested by users in their computers. Only the best graphics card deserves the highest score in the 3DMark Vantage Performance benchmark.

Passmark GPU

Passmark is an effective benchmark that displays true results on the speed of data processing by GPU cores. We present the results of the evaluation of the performance of a video card in the modern Passmark benchmark, which evaluates the performance of a video card using its own rendering algorithms. The higher the score, the more powerful the graphics card.

GeekBench 5 CUDA

Detailed benchmark results for GeekBench 5 CUDA video cards in the list of all video cards can be found on this page. The higher the GeekBench 5 score, the faster and more powerful the graphics card.

GeekBench 5 OpenCL

Benchmark GeekBench 5 OpenCL is a specialized version of a synthetic test that determines how the technical characteristics of a video card cope with loads in different programmable scenarios. Based on version 5, updated regularly.

SPECviewperf 12 - specvp12 sw-03

The sw-03 viewset was created from traces of Dassault Systemes’ SolidWorks 2013 SP1 application. Models used in the viewset range in size from 2.1 to 21 million vertices.

SPECviewperf 12 - specvp12 snx-02

The snx-02 viewset was created from traces of the graphics workload generated by the NX 8.0 application from Siemens PLM. Model sizes range from 7.15 to 8.45 million vertices. The viewset includes numerous rendering modes supported by the application, including wireframe, anti-aliasing, shaded, shaded with edges, and studio mode.

SPECviewperf 12 - specvp12 showcase-01

The showcase-01 view set in the SPECviewperf 12 benchmark is based on Autodesk Showcase 2013. Supports DX rendering. Tests video cards in modes - shading with own shadows and projection shadows.

SPECviewperf 12 - specvp12 medical-01

The SPECviewperf 12 benchmark in specvp12 medical-01 script mode renders a 2D projection while calculating texturing and further GPU performance calculations. Works in scenes on 4D heart models and 3D stag beetle models.

SPECviewperf 12 - specvp12 maya-04

The maya-04 viewset was created from traces of the graphics workload generated by the Maya 2013 application from Autodesk. Model size is 727,500 vertices. The viewset includes numerous rendering modes supported by the application, including shaded mode, ambient occlusion, multi-sample antialiasing, and transparency.

SPECviewperf 12 - specvp12 energy-01

The SPECviewperf 12 benchmark version of the specvp12 energy-01 script tests the volumetric rendering mode in the video card in seismic and oil and gas fields. This is a rather complicated benchmark in terms of the load on the computing core of the video card.

SPECviewperf 12 - specvp12 creo-01

Benchmark SPECviewperf 12 version of the "specvp12 creo-01" script tests the video card performance for the following render models - "supercar" and "airplane". Powered by Creo 2™ from PTC. Creates various loads on the GPU core to evaluate the efficiency and speed of the video card.

SPECviewperf 12 - specvp12 catia-04

The snx-02 viewset was created from traces of the graphics workload generated by the NX 8.0 application from Siemens PLM. Model sizes range from 7.15 to 8.45 million vertices. The viewset includes numerous rendering modes supported by the application, including wireframe, anti-aliasing, shaded, shaded with edges, and studio mode.

SPECviewperf 12 - specvp12 3dsmax-05

The 3dsmax-05 viewset was created from traces of the graphics workload generated by 3ds Max 2016 using the default Nitrous DX11 driver.

GeekBench 5 Vulkan

Benchmark GeekBench 5 Vulkan shows the result of testing video cards in a specially created 3D environment. Results are updated regularly, daily.

5.0 Out of 1 CMPTEK score