GeForce RTX 5090 vs RTX 3090
VS
Nvidia GeForce RTX 5090
GB202-300
Nvidia GeForce RTX 3090
GA102
We compared two discrete desktop gaming GPUs: the GeForce RTX 5090 32 GB with 170 pipelines and 21760 shaders against the 4 years and 4 months older RTX RTX 3090 24 GB that utilizes 82 pipelines and 10496 shaders. Here you will find complete details about specs, efficiency, performance tests, and more.
Review
Gaming
Performance in DirectX, OpenCL, and Vulkan games
Workstation
Perf. in 3D modeling, video editing and rendering apps
AI/ML
Capabilities for machine learning and AI-related tasks
Energy Efficiency
Power consumption efficiency in different scenarios
NanoReview Final Score
Overall video card score
Value for money (Beta)
Key differences
Reasons to consider the GeForce RTX 5090
- Performs significantly better (up to 2.4x) in 3DMark Steel Nomad Lite
- 2.9x higher maximum theoretical performance (104.8 vs 35.6 TFLOPS)
- Manufactured using a more efficient 4 nm process technology
- Shows 73% higher average frame rate in modern games at QHD resolution – 207 vs 120 FPS
Gaming Performance
Frame rate comparison across popular AAA titles at different resolutionsGames
| Forza Horizon 5 | 336 247 233 187 |
186 128 117 91 |
| The Witcher 3 | 364 327 237 123 |
205 170 128 66 |
| Counter-Strike 2 | 363 265 202 99 |
275 207 155 83 |
| Far Cry 6 | 237 202 190 121 |
156 145 124 77 |
| Hogwarts Legacy | 249 199 166 97 |
128 108 86 47 |
| CoD: Modern Warfare III | 334 322 260 184 |
171 159 123 80 |
| Ghost of Tsushima | 221 185 175 113 |
120 97 85 48 |
| Cyberpunk 2077 | 261 233 160 77 |
148 132 90 43 |
| Shadow of the Tomb Raider | 329 302 238 128 |
237 226 168 91 |
| 1080p High | 299 | 181 |
| 1080p Ultra | 254 | 152 |
| 1440p Ultra | 207 | 120 |
| 4K Ultra | 125 | 70 |
| Margin of Error | Low | Low |
Benchmarks
Graphics cards’ performance in recent benchmarking apps3D Mark
Steel Nomad Lite Score
| Time Spy | 46873 | 19896 |
| Solar Bay | 236689 | 97395 |
| Port Royal | 38202 | 13633 |
| Fire Strike | 87569 | 47512 |
| Wild Life Extreme | 108751 | 43669 |
| Night Raid | 206648 | 139766 |
GeekBench 6 OpenCL
GB6 Compute Score
| Background Blur | 333 img/sec | 320.8 img/sec |
| Face Detection | 298.5 img/sec | 207.2 img/sec |
| Horizon Detection | 21.7 Gpixels/sec | 10.1 Gpixels/sec |
| Edge Detection | 33.7 Gpixels/sec | 14.4 Gpixels/sec |
| Gaussian Blur | 34.5 Gpixels/sec | 11.5 Gpixels/sec |
| Feature Matching | 2.53 Gpixels/sec | 1.48 Gpixels/sec |
| Stereo Matching | 2850 Gpixels/sec | 984.2 Gpixels/sec |
| Particle Physics | 59146.9 FPS | 27764.9 FPS |
| API | OpenCL | OpenCL |
Cinebench 2024 GPU
Cinebench 2024 GPU
Passmark Graphics
G3D Mark Score
| G2D Mark | 1410 | 1069 |
| DirectX 11 | 340 FPS | 218 FPS |
| DirectX 12 | 179 FPS | 109 FPS |
| GPU Compute | 24383 Ops/s | 15028 Ops/s |
Blender
Blender GPU
Artificial Intelligence Tests
Performance in machine learning and artificial intelligence tasksGeekBench 6 ML
GB6 ML Single Precision
GB6 ML Half Precision
GB6 ML Quantized
| Image Classification (SP) | 18948 | 10665 |
| Image Segmentation (HP) | 55109 | 28963 |
| Image Super Resolution (Q) | 58478 | 22697 |
| Face Detection (HP) | 114685 | 55232 |
| Pose Estimation (Q) | 344729 | 128155 |
| Text Classification (SP) | 4464 | 3152 |
| Machine Translation (HP) | 8184 | 5088 |
| Object Detection (SP) | 27802 | 13891 |
| Depth Estimation (Q) | 85572 | 37984 |
| Style Transfer (SP) | 687080 | 272000 |
| Framework | ONNX | ONNX |
| Backend | DirectML | DirectML |
Specifications
Technical specifications of GeForce RTX 5090 and RTX 3090General
| Vendor | Nvidia | Nvidia |
| Build | Discrete | Discrete |
| Released | January 7, 2025 | September 24, 2020 |
| Launch price (MSRP) | $1999 | $1499 |
| Case | Desktop | Desktop |
| Purpose | Gaming | Gaming |
| Segment | High-end | High-end |
| Architecture | Blackwell 2.0 | Ampere |
| GPU Codename | GB202-300 | GA102 |
| Rival Equivalent | - | - Radeon RX 6900 XT |
| Successor | - | - GeForce RTX 5090 |
| Recommended CPU | - Intel Core Ultra 9 285K or above | - Intel Core i9 12900K or above |
Desktop GPU rating (2nd and 20th place)
Graphics Processing Unit
| Base Clock | 2017 MHz | 1395 MHz |
| Boost Clock | 2407 MHz | 1695 MHz |
| Shading Units | 21760 | 10496 |
| Texture Mapping Units (TMUs) | 680 | 328 |
| Render Output Units (ROPs) | 176 | 112 |
| Compute Units (Pipelines) | 170 | 82 |
| Tensor Cores | 680 | 328 |
| Ray-tracing Cores | 170 | 82 |
| L1 Cache | 128KB per cluster | 128KB per cluster |
| L2 Cache | 96MB shared | 6MB shared |
| Instructions Per Cycle | 2 IPC | 2 IPC |
Raw Performance
| Pixel Fill Rate | 424 GPixel/s | 190 GPixel/s |
| Texture Fill Rate | 1637 GTexel/s | 556 GTexel/s |
FLOPS (FP32)
Physical
| Interface | PCIe 5.0 x16 | PCIe 4.0 x16 |
| TGP | 575 W | 350 W |
| Manufacturing | TSMC | Samsung |
| Fabrication Process | 4 nm | 8 nm |
| Die Size | 750 mm² | 628 mm² |
| Transistor Count | 92.2 billion | 28 billion |
| Transistor Density | 122.93 MTr/mm² | 44.59 MTr/mm² |
Memory
| Memory Type | GDDR7 | GDDR6X |
| Memory Size | 32 GB | 24 GB |
| Memory Clock | 1750 MHz | 1219 MHz |
| Effective Memory Speed | 28000 Mbps | 19500 Mbps |
| Bus | 512-bit | 384-bit |
| ECC | No | No |
Memory Bandwidth
API
| DirectX | 12.2 | 12 |
| Vulkan | 1.4 | 1.3 |
| OpenGL | 4.6 | 4.6 |
| OpenCL | 3.0 | 3.0 |
| CUDA | 12.0 | 8.6 |
| Ray Tracing | Yes | Yes |
| DLSS | DLSS 4 | DLSS 2 |
| DisplayPort | 2.1b | 1.4a |
Cast your vote
Total votes: 66
