ASUS ROG Strix GeForce RTX 4090
vs
ASUS ROG Strix GeForce RTX 3090

GPU comparison with benchmarks

Description

In the gpu comparison between the ASUS ROG Strix GeForce RTX 4090 and the ASUS ROG Strix GeForce RTX 3090, we compare both the technical data and the benchmark results of the two graphics cards.

The ASUS ROG Strix GeForce RTX 4090 is based on the NVIDIA GeForce RTX 4090 (Ada Lovelace) and has 24 GB GDDR6X graphics memory with a bandwidth of 1008 GB/s. The 16384 shader units clock with a maximum of 2.55 GHz which results in an FP32 processing power of 83.56 TFLOPS.

The ASUS ROG Strix GeForce RTX 3090 is based on the NVIDIA GeForce RTX 3090 (Ampere) and has 24 GB GDDR6X graphics memory with a bandwidth of 936 GB/s. The 10496 shader units have a maximum clock rate of 1.73 GHz and achieve an FP32 computing power of 36.21 TFLOPS.

GPU

The GPU core data includes the number of execution units and texture shaders. Both have a major impact on the computing power of a graphics card.
ASUS ROG Strix GeForce RTX 4090 GPU ASUS ROG Strix GeForce RTX 3090
NVIDIA GeForce RTX 4090 Based on NVIDIA GeForce RTX 3090
AD102 GPU Chip GA102-300-A1
Ada Lovelace Architecture Ampere
128 Execution units 82
16384 Shader 10496
192 Render Output Units 112
512 Texture Units 328
128 Raytracing Cores 82

Memory

A large graphics memory is important to process large and detailed textures quickly, even in high resolutions. The speed of the memory and the resulting memory bandwidth are also important for this.

24 GB Memory Size 24 GB
GDDR6X Memory Type GDDR6X
1.313 GHz Memory Clock 1.219 GHz
21 Gbps Memory Speed 19.5 Gbps
1008 GB/s Memory Bandwith 936 GB/s
384 bit Memory Interface 384 bit

Clock Speeds

The clock frequency of a graphics card determines the speed in connection with the graphics architecture. New manufacturing processes are making ever higher clock frequencies possible, whereby attention must also be paid to the energy consumption of the graphics card, as this also depends to a large extent on the clock frequency.

2.235 GHz Base Clock 1.395 GHz
2.550 GHz Boost Clock 1.725 GHz
Avg (Game) Clock 1.70 GHz
Yes Overclocking Yes

Thermal Design

The TDP defines the maximum power consumption of a graphics card. The ventilation of the computer and the performance of the power supply are based on the TDP and should be optimized for it.

450 W TDP 350 W
450 W TDP (up) 430 W
90 °C Tjunction max 93 °C
1 x 16-Pin PCIe-Power 3 x 8-Pin

Cooler & Fans

The cooler type as well as the number and size of the fans of a graphics card result in its volume. A loud graphics card can easily be perceived as annoying. Generally, larger fans are better because they can spin slower to move the same volume of air as opposed to smaller fans.

Axial Fan-Type Axial
3 x 110 mm Fan 1 3 x 95 mm
-- Fan 2 --
Air cooling Cooler-Type Air cooling
0 dB Noise (Idle) 0 dB
-- Noise (Load) 38-45 dB

Connectivity

In the connectivity category, we recorded the number of ports on a graphics card. You can also find out here the maximum number of screens you can connect to the graphics card.

4 Max. Displays 4
2.3 HDCP-Version 2.3
2x HDMI v2.1a HDMI Ports 2x HDMI v2.1
3x DP v1.4a DP Ports 3x DP v1.4a
-- DVI Ports --
-- VGA Ports --
-- USB-C Ports --

Featureset

Here we show you which functions a graphics card has. We've also included ray tracing support for true simulation of light rays and shadows.

7680x4320 Max. resolution 7680x4320
12_2 (Ultimate) DirectX 12_2 (Ultimate)
Yes Raytracing Yes
Yes DLSS / FSR Yes
ASUS Aura Sync LED ASUS Aura Sync

Supported Video Codecs

The support and acceleration of certain video codecs directly in the graphics card reduces the load when playing or creating videos. This can have a positive impact on the lower energy consumption and less waste heat from the graphics card.

Decode / Encode h264 Decode / Encode
Decode / Encode h265 / HEVC Decode / Encode
Decode AV1 Decode
Decode VP8 Decode
Decode VP9 Decode

Dimensions

With the exact dimensions of a graphics card you can check whether it fits into your computer case. We also show you the PCI Express standard that the graphics card supports at most.

358 mm Length 319 mm
149 mm Height 140 mm
70 mm Width 58 mm
4 PCIe-Slots Width (Slots) 3 PCIe-Slots
-- Weight 1810 g
PCIe 4.0 x 16 GPU Interface PCIe 4.0 x 16

Additional data

In addition to data such as the release date and the price when the graphics card is released, we usually show you a link to the data sheet for the graphics card directly from the manufacturer.

ROG-STRIX-RTX4090-24G-GAMING Part-no ROG-STRIX-RTX3090-24G-GAMING
Q4/2022 Release date Q4/2020
1,599 $ (Reference) Launch Price 1,499 $ (Reference)
4 nm Structure size 8 nm
data sheet Documents data sheet

Benchmark results

FP32 Performance (Single-precision TFLOPS) - FP32 (TFLOPS)

The theoretical computing power of the graphics card with single precision (32 bit) in TFLOPS. TFLOPS indicates how many trillion FP32 floating point operations the graphics card (GPU) can perform per second. The larger this number, the faster the graphics card is.




3DMark Benchmark (DirectX, Raytracing) - Time Spy Graphics score

3DMark is a benchmark program that determines the performance of certain components of a computer and then reports the performance as a numerical value. Here at GPU-Monkey we only use selected benchmarks for the graphics card, which are determined with a sufficiently fast processor so that the processor does not negatively affect the performance of the graphics card.




3DMark Benchmark (DirectX, Raytracing) - Time Spy Extreme Graphics score

3DMark is a benchmark program that determines the performance of certain components of a computer and then reports the performance as a numerical value. Here at GPU-Monkey we only use selected benchmarks for the graphics card, which are determined with a sufficiently fast processor so that the processor does not negatively affect the performance of the graphics card.




3DMark Benchmark (DirectX, Raytracing) - Fire Strike Extreme Graphics score

3DMark is a benchmark program that determines the performance of certain components of a computer and then reports the performance as a numerical value. Here at GPU-Monkey we only use selected benchmarks for the graphics card, which are determined with a sufficiently fast processor so that the processor does not negatively affect the performance of the graphics card.




3DMark Benchmark (DirectX, Raytracing) - Port Royal (Raytracing)

3DMark is a benchmark program that determines the performance of certain components of a computer and then reports the performance as a numerical value. Here at GPU-Monkey we only use selected benchmarks for the graphics card, which are determined with a sufficiently fast processor so that the processor does not negatively affect the performance of the graphics card.




3DMark Benchmark (DirectX, Raytracing) - Speed Way Graphics Score (Raytracing)

3DMark is a benchmark program that determines the performance of certain components of a computer and then reports the performance as a numerical value. Here at GPU-Monkey we only use selected benchmarks for the graphics card, which are determined with a sufficiently fast processor so that the processor does not negatively affect the performance of the graphics card.




Geekbench 5 (CUDA, OpenCL, Vulkan, Metal) - Geekbench 5 - CUDA (NVIDIA only)

Geekbench 5 is primarily a cross-platform benchmark for main processors, which runs through various tests in order to finally output the processor's performance in the form of a numerical value. In addition to the CPU benchmark, Geekbench can also run 4 different graphics benchmarks and output them in the form of a numerical value.
Zusätzlich zum CPU-Benchmark kann Geekbench auch 4 verschiedene Grafik-Benchmarks durchführen und in Form eines Zahlenwertes ausgeben.




Geekbench 5 (CUDA, OpenCL, Vulkan, Metal) - Geekbench 5 - OpenCL

Geekbench 5 is primarily a cross-platform benchmark for main processors, which runs through various tests in order to finally output the processor's performance in the form of a numerical value. In addition to the CPU benchmark, Geekbench can also run 4 different graphics benchmarks and output them in the form of a numerical value.
Zusätzlich zum CPU-Benchmark kann Geekbench auch 4 verschiedene Grafik-Benchmarks durchführen und in Form eines Zahlenwertes ausgeben.




Geekbench 5 (CUDA, OpenCL, Vulkan, Metal) - Geekbench 5 - Vulkan

Geekbench 5 is primarily a cross-platform benchmark for main processors, which runs through various tests in order to finally output the processor's performance in the form of a numerical value. In addition to the CPU benchmark, Geekbench can also run 4 different graphics benchmarks and output them in the form of a numerical value.
Zusätzlich zum CPU-Benchmark kann Geekbench auch 4 verschiedene Grafik-Benchmarks durchführen und in Form eines Zahlenwertes ausgeben.




Cyberpunk 2077 - 3840x2160 (2160p)

Cyberpunk 2077 is an open-world action role-playing game developed by CD Projekt Red and based on the development studio's in-house REDengine 4. In our Cyberpunkt 2077 benchmarks, ray tracing is turned off.




Cyberpunk 2077 - 2560x1440 (1440p)

Cyberpunk 2077 is an open-world action role-playing game developed by CD Projekt Red and based on the development studio's in-house REDengine 4. In our Cyberpunkt 2077 benchmarks, ray tracing is turned off.




Cyberpunk 2077 - 1920x1080 (1080p)

Cyberpunk 2077 is an open-world action role-playing game developed by CD Projekt Red and based on the development studio's in-house REDengine 4. In our Cyberpunkt 2077 benchmarks, ray tracing is turned off.




Battlefield 5 - 3840x2160 (2160p)

Battlefield 5 is a visually impressive game. It uses the third version of the "Frostbite Game Engine" developed by the manufacturer DICE. The engine is available for Windows, PS4, PS5, XBox One or XBox Series X. Many physics-based calculations are carried out directly in the game engine, which results in very realistic calculations. Due to its current features, Battlefield 5 is very suitable as a graphics card benchmark.

Test environment and settings: DirectX 12 API with max. Details, tested under Windows 10.




Battlefield 5 - 2560x1440 (1440p)

Battlefield 5 is a visually impressive game. It uses the third version of the "Frostbite Game Engine" developed by the manufacturer DICE. The engine is available for Windows, PS4, PS5, XBox One or XBox Series X. Many physics-based calculations are carried out directly in the game engine, which results in very realistic calculations. Due to its current features, Battlefield 5 is very suitable as a graphics card benchmark.

Test environment and settings: DirectX 12 API with max. Details, tested under Windows 10.




Battlefield 5 - 1920x1080 (1080p)

Battlefield 5 is a visually impressive game. It uses the third version of the "Frostbite Game Engine" developed by the manufacturer DICE. The engine is available for Windows, PS4, PS5, XBox One or XBox Series X. Many physics-based calculations are carried out directly in the game engine, which results in very realistic calculations. Due to its current features, Battlefield 5 is very suitable as a graphics card benchmark.

Test environment and settings: DirectX 12 API with max. Details, tested under Windows 10.




Shadow of the Tomb Raider - 3840x2160 (2160p)

Shadow of the Tomb Raider is based on a modified version of the Crystal Dynamics Foundation engine. This engine was already used in the predecessor Rise of the Tomb Raider. The foundation engine supports a realistic representation of water and snow and shows the game characters detailed. The display of hair is also very realistic thanks to the support of TressFX (AMD) or HairWorks (NVIDIA).




Shadow of the Tomb Raider - 2560x1440 (1440p)

Shadow of the Tomb Raider is based on a modified version of the Crystal Dynamics Foundation engine. This engine was already used in the predecessor Rise of the Tomb Raider. The foundation engine supports a realistic representation of water and snow and shows the game characters detailed. The display of hair is also very realistic thanks to the support of TressFX (AMD) or HairWorks (NVIDIA).




Shadow of the Tomb Raider - 1920x1080 (1080p)

Shadow of the Tomb Raider is based on a modified version of the Crystal Dynamics Foundation engine. This engine was already used in the predecessor Rise of the Tomb Raider. The foundation engine supports a realistic representation of water and snow and shows the game characters detailed. The display of hair is also very realistic thanks to the support of TressFX (AMD) or HairWorks (NVIDIA).




GTA 5 Benchmark (Grand Theft Auto V) - 3840x2160 (2160p)

Grand Theft Auto V (GTA 5) is based on the RAGE Engine and was developed by Rockstar North. Rockstar had already released the open-world game for the Xbox and Playstation consoles at the end of 2013, and the PC version followed in spring 2015.




GTA 5 Benchmark (Grand Theft Auto V) - 2560x1440 (1440p)

Grand Theft Auto V (GTA 5) is based on the RAGE Engine and was developed by Rockstar North. Rockstar had already released the open-world game for the Xbox and Playstation consoles at the end of 2013, and the PC version followed in spring 2015.




GTA 5 Benchmark (Grand Theft Auto V) - 1920x1080 (1080p)

Grand Theft Auto V (GTA 5) is based on the RAGE Engine and was developed by Rockstar North. Rockstar had already released the open-world game for the Xbox and Playstation consoles at the end of 2013, and the PC version followed in spring 2015.




Crypto-Mining Ethereum Hashrate (MH/s) - Ethash Hashrate

Ethash is the Proof-of-Work (PoW) algorithm used for the crypto currency Ethereum. Ethash is based on the Dagger-Hashimoto algorithm, but has been significantly changed in some places. For example, the memory requirement (GPU memory) was significantly increased in order to make it more difficult or even to prevent the calculation of the algorithm on so-called ASICs (Application Specific Integrated Circuits). Our benchmark results are based on a slight overclocking of the graphics card and the GPU memory. The hashrates were achieved under Hive OS with moderate overclocking settings.




Crypto-Mining Ergo Hashrate (MH/s) - Autolykos 2 Hashrate

The crypto currency Ergo is based on the Autolykos algorithm, which was updated in version 2 (Autolykos 2) in early 2021. The Autolykos 2 algorithm is based on the PoW concept and can be efficiently calculated by a graphics card. When calculating Ergo Coins, ASICs have no efficiency advantage over modern graphics cards, which should make Ergo attractive for as many miners as possible. The Ergo protocol is based on the blockchain and has been optimized for a very long shelf life. The hashrates were achieved under Hive OS with moderate overclocking settings.




Crypto-Mining Ravencoin Hashrate (MH/s) - KawPow Hashrate

The KawPoW algorithm has been used with the crypto currency Ravencoin since May 2020. Previously, Ravencoin used ProgPoW (X16R and X16Rv2). KawPoW is a Proof-of-Work (PoW) algorithm and is calculated the fastest with a modern graphics card. KawPoW is based on a changing algorithm (X15 and SHA51) and can therefore not be calculated via ASIC. The developers want to protect Ravencoin from large mining farms that often use ASICs. In addition to Ravencoin, other cryptocurrencies are based on KawPoW, including Hilux and Gravium. The hashrates were achieved under Hive OS with moderate overclocking settings.




Crypto-Mining Vertcoin Hashrate (MH/s) - Verthash Hashrate

The crypto currency VertCoin (VTC) is calculated with the new Verthash algorithm since January 2021. This is a very memory-intensive PoW algorithm that is difficult to calculate with an ASIC. The developers have set themselves the goal that VertCoin should be calculated via a graphics card and not with so-called ASICs. These are increasingly used in large mining farms. Verthash currently requires a graphics card with at least 2 GB of memory. The hashrates were achieved under Hive OS with moderate overclocking settings.




Popular comparisons

1. ASUS ROG Strix GeForce RTX 4090 ASUS TUF Gaming GeForce RTX 4090 ASUS ROG Strix GeForce RTX 4090 vs ASUS TUF Gaming GeForce RTX 4090
2. ASUS ROG Strix GeForce RTX 3090 MSI GeForce RTX 3090 Gaming X Trio 24G ASUS ROG Strix GeForce RTX 3090 vs MSI GeForce RTX 3090 Gaming X Trio 24G
3. ASUS ROG Strix GeForce RTX 3090 ASUS ROG Strix GeForce RTX 3090 OC ASUS ROG Strix GeForce RTX 3090 vs ASUS ROG Strix GeForce RTX 3090 OC
4. NVIDIA GeForce RTX 4090 Founders Edition ASUS ROG Strix GeForce RTX 4090 NVIDIA GeForce RTX 4090 Founders Edition vs ASUS ROG Strix GeForce RTX 4090
5. ASUS ROG Strix GeForce RTX 4090 OC Edition ASUS ROG Strix GeForce RTX 4090 ASUS ROG Strix GeForce RTX 4090 OC Edition vs ASUS ROG Strix GeForce RTX 4090
6. MSI GeForce RTX 4090 SUPRIM LIQUID X 24G ASUS ROG Strix GeForce RTX 4090 MSI GeForce RTX 4090 SUPRIM LIQUID X 24G vs ASUS ROG Strix GeForce RTX 4090
7. ASUS ROG Strix GeForce RTX 4090 ZOTAC GAMING GeForce RTX 4090 Trinity OC ASUS ROG Strix GeForce RTX 4090 vs ZOTAC GAMING GeForce RTX 4090 Trinity OC
8. ASUS ROG Strix GeForce RTX 3090 ASUS TUF Gaming GeForce RTX 3090 OC ASUS ROG Strix GeForce RTX 3090 vs ASUS TUF Gaming GeForce RTX 3090 OC
9. ASUS ROG Strix GeForce RTX 3090 ZOTAC Gaming GeForce RTX 3090 Trinity OC ASUS ROG Strix GeForce RTX 3090 vs ZOTAC Gaming GeForce RTX 3090 Trinity OC
10. ASUS ROG Strix GeForce RTX 3090 ASUS ROG Strix LC Radeon RX 6900 XT OC ASUS ROG Strix GeForce RTX 3090 vs ASUS ROG Strix LC Radeon RX 6900 XT OC


back to index
Facebook
Twitter