NVIDIA GeForce RTX 4090 vs NVIDIA GeForce RTX 3090

Comparative analysis of NVIDIA GeForce RTX 4090 and NVIDIA GeForce RTX 3090 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, 3DMark Fire Strike - Graphics Score, CompuBench 1.5 Desktop - Face Detection (mPixels/s), CompuBench 1.5 Desktop - T-Rex (Frames/s), CompuBench 1.5 Desktop - Video Composition (Frames/s), CompuBench 1.5 Desktop - Bitcoin Mining (mHash/s), Geekbench - OpenCL, CompuBench 1.5 Desktop - Ocean Surface Simulation (Frames/s), GFXBench 4.0 - Car Chase Offscreen (Frames), GFXBench 4.0 - Manhattan (Frames), GFXBench 4.0 - T-Rex (Frames), GFXBench 4.0 - Car Chase Offscreen (Fps), GFXBench 4.0 - Manhattan (Fps), GFXBench 4.0 - T-Rex (Fps).

NVIDIA GeForce RTX 4090

NVIDIA GeForce RTX 3090

Differences

Reasons to consider the NVIDIA GeForce RTX 4090

Videocard is newer: launch date 2 year(s) 0 month(s) later
Around 60% higher core clock speed: 2235 MHz vs 1395 MHz
Around 49% higher boost clock speed: 2520 MHz vs 1695 MHz
Around 56% higher pipelines: 16384 vs 10496
A newer manufacturing process allows for a more powerful, yet cooler running videocard: 4 nm vs 8 nm
Around 8% higher memory clock speed: 1313 MHz, 21 Gbps effective vs 1219 MHz (19.5 Gbps effective)
Around 22% better performance in PassMark - G2D Mark: 1297 vs 1060
Around 44% better performance in PassMark - G3D Mark: 38287 vs 26668
Around 80% better performance in 3DMark Fire Strike - Graphics Score: 9223 vs 5118
Around 57% better performance in CompuBench 1.5 Desktop - T-Rex (Frames/s): 98.472 vs 62.614
Around 22% better performance in CompuBench 1.5 Desktop - Bitcoin Mining (mHash/s): 2968.159 vs 2441.384
Around 66% better performance in Geekbench - OpenCL: 317791 vs 191142

Specifications (specs)
Launch date	20 Sep 2022 vs 1 Sep 2020
Core clock speed	2235 MHz vs 1395 MHz
Boost clock speed	2520 MHz vs 1695 MHz
Pipelines	16384 vs 10496
Manufacturing process technology	4 nm vs 8 nm
Memory clock speed	1313 MHz, 21 Gbps effective vs 1219 MHz (19.5 Gbps effective)
Benchmarks
PassMark - G2D Mark	1297 vs 1060
PassMark - G3D Mark	38287 vs 26668
3DMark Fire Strike - Graphics Score	9223 vs 5118
CompuBench 1.5 Desktop - T-Rex (Frames/s)	98.472 vs 62.614
CompuBench 1.5 Desktop - Bitcoin Mining (mHash/s)	2968.159 vs 2441.384
Geekbench - OpenCL	317791 vs 191142

Reasons to consider the NVIDIA GeForce RTX 3090

431x more texture fill rate: 556.0 GTexel/s vs 1,290 GTexel/s
Around 29% lower typical power consumption: 350 Watt vs 450 Watt
Around 19% better performance in CompuBench 1.5 Desktop - Face Detection (mPixels/s): 690.619 vs 582.642
Around 17% better performance in CompuBench 1.5 Desktop - Video Composition (Frames/s): 209.424 vs 178.756

Specifications (specs)
Texture fill rate	556.0 GTexel/s vs 1,290 GTexel/s
Thermal Design Power (TDP)	350 Watt vs 450 Watt
Benchmarks
CompuBench 1.5 Desktop - Face Detection (mPixels/s)	690.619 vs 582.642
CompuBench 1.5 Desktop - Video Composition (Frames/s)	209.424 vs 178.756

Compare benchmarks

GPU 1: NVIDIA GeForce RTX 4090
GPU 2: NVIDIA GeForce RTX 3090

PassMark - G2D Mark

GPU 1

GPU 2

1297

1060

PassMark - G3D Mark

GPU 1

GPU 2

38287

26668

3DMark Fire Strike - Graphics Score

GPU 1

GPU 2

9223

5118

CompuBench 1.5 Desktop - Face Detection (mPixels/s)

GPU 1

GPU 2

582.642

690.619

CompuBench 1.5 Desktop - T-Rex (Frames/s)

GPU 1

GPU 2

98.472

62.614

CompuBench 1.5 Desktop - Video Composition (Frames/s)

GPU 1

GPU 2

178.756

209.424

CompuBench 1.5 Desktop - Bitcoin Mining (mHash/s)

GPU 1

GPU 2

2968.159

2441.384

Geekbench - OpenCL

GPU 1

GPU 2

317791

191142

Name	NVIDIA GeForce RTX 4090	NVIDIA GeForce RTX 3090
PassMark - G2D Mark	1297	1060
PassMark - G3D Mark	38287	26668
3DMark Fire Strike - Graphics Score	9223	5118
CompuBench 1.5 Desktop - Face Detection (mPixels/s)	582.642	690.619
CompuBench 1.5 Desktop - T-Rex (Frames/s)	98.472	62.614
CompuBench 1.5 Desktop - Video Composition (Frames/s)	178.756	209.424
CompuBench 1.5 Desktop - Bitcoin Mining (mHash/s)	2968.159	2441.384
Geekbench - OpenCL	317791	191142
CompuBench 1.5 Desktop - Ocean Surface Simulation (Frames/s)	0	7585.258
GFXBench 4.0 - Car Chase Offscreen (Frames)		33398
GFXBench 4.0 - Manhattan (Frames)		3713
GFXBench 4.0 - T-Rex (Frames)		3354
GFXBench 4.0 - Car Chase Offscreen (Fps)		33398
GFXBench 4.0 - Manhattan (Fps)		3713
GFXBench 4.0 - T-Rex (Fps)		3354

Compare specifications (specs)

	NVIDIA GeForce RTX 4090	NVIDIA GeForce RTX 3090
Essentials
Architecture	Ada Lovelace	Ampere
Code name	AD102	GA102
Launch date	20 Sep 2022	1 Sep 2020
Place in performance rating	13	46
Launch price (MSRP)		$1499
Type		Desktop
Technical info
Boost clock speed	2520 MHz	1695 MHz
Core clock speed	2235 MHz	1395 MHz
Manufacturing process technology	4 nm	8 nm
Peak Double Precision (FP64) Performance	1,290 GFLOPS (1:64)	556.0 GFLOPS (1:64)
Peak Half Precision (FP16) Performance	82.58 TFLOPS (1:1)	35.58 TFLOPS (1:1)
Peak Single Precision (FP32) Performance	82.58 TFLOPS	35.58 TFLOPS
Pipelines	16384	10496
Pixel fill rate	443.5 GPixel/s	189.8 GPixel/s
Texture fill rate	1,290 GTexel/s	556.0 GTexel/s
Thermal Design Power (TDP)	450 Watt	350 Watt
Transistor count	76300 million	28300 million
Video outputs and ports
Display Connectors	1x HDMI 2.1, 3x DisplayPort 1.4a	1x HDMI, 3x DisplayPort
Compatibility, dimensions and requirements
Form factor	Triple-slot
Height	61 mm, 2.4 inches	138 mm (5.4 inches)
Interface	PCIe 4.0 x16	PCIe 4.0 x16
Length	304 mm, 12 inches	313 mm (12.3 inches)
Recommended system power (PSU)	850 Watt	750 Watt
Supplementary power connectors	1x 16-pin	1x 12-pin
Width	137 mm, 5.4 inches	Triple-slot
API support
DirectX	12 Ultimate (12_2)	12.2
OpenCL	3.0	2.0
OpenGL	4.6	4.6
Shader Model	6.7	6.5
Vulkan
Memory
Maximum RAM amount	24 GB	24 GB
Memory bandwidth	1,008 GB/s	936.2 GB/s
Memory bus width	384 bit	384 bit
Memory clock speed	1313 MHz, 21 Gbps effective	1219 MHz (19.5 Gbps effective)
Memory type	GDDR6X	GDDR6X