Nvidia GeForce RTX 5090 vs Nvidia GeForce RTX 3090 Ti
97
out 100
VS
71
out 100
Compared 2 processors: Nvidia GeForce RTX 5090, vs Nvidia GeForce RTX 3090 Ti. Below are specs, tests, pros and cons for each device
Review
Games
Performance in DirectX, OpenCL, Vulkan
AI/ML
AI workload testing
Professional Apps
3D modeling, video editing, rendering
Energy Efficiency
Power usage in various scenarios
Tests & specs
Specs & test comparisonGames
| Game title | 1080p (High) 1080p (Ultra) 1440p (Ultra) 4K (Ultra) | |
 Forza Horizon 5 | 332 242 231 183 | 196 139 127 96 |
 The Witcher 3 | 366 338 234 127 | 218 185 131 70 |
 Counter-Strike 2 | 379 271 209 98 | 290 218 161 86 |
 Far Cry 6 | 235 202 198 124 | 158 148 129 81 |
 Hogwarts Legacy | 244 210 167 96 | 136 113 91 50 |
 COD Modern Warfare 3 | 337 333 269 184 | 181 173 136 89 |
 Ghost of Tsushima | 221 190 170 114 | 124 106 91 55 |
 Cyberpunk 2077 | 259 231 166 77 | 163 146 98 45 |
 Shadow of the Tomb Raider | 334 306 238 128 | 261 238 186 101 |
| 1080p High | 301 | 192 |
| 1080p Ultra | 258 | 163 |
| 1440p Ultra | 209 | 128 |
| 4K Ultra | 126 | 75 |
| Deviation | Low | Low |
Benchmarks
| Time Spy | 47515 | 21841 |
| Solar Bay | 213229 | 109151 |
| Port Royal | 37529 | 14882 |
| Fire Strike | 90200 | 51985 |
| Night Raid | 209016 | 155722 |
| Background Blur | 333 frames/s | 362.7 frames/s |
| Face Detection | 298.5 frames/s | 227.9 frames/s |
| Horizon Detection | 21.7 Gpix/sec | 11.3 Gpix/sec |
| Edge Detection | 33.7 Gpix/sec | 15.9 Gpix/sec |
| Gaussian Blur | 34.5 Gpix/sec | 12.1 Gpix/sec |
| Feature Matching | 2.53 Gpix/sec | 1.56 Gpix/sec |
| Stereo Matching | 2850 Gpix/sec | 1080 Gpix/sec |
| Particle Physics | 59146.9 fps | 31024.6 fps |
| API | OpenCL | OpenCL |
| G2D Mark | 1444 | 1227 |
| DirectX 11 | 344 fps | 241 fps |
| DirectX 12 | 195 fps | 122 fps |
| GPU Compute | 28243 Ops/s | 17022 Ops/s |
| Aztec Ruins (Normal) | 1397.2 fps | 1005.1 fps |
| Car Chase | 1150.3 fps | 897.2 fps |
| Manhattan 3.1.1 (1440p) | 1178.2 fps | 781.7 fps |
| T-Rex | 4812.3 fps | 4109.9 fps |
| ALU 2 | 5992.8 fps | 7102.5 fps |
| Driver Overhead 2 | 269.7 fps | 274.9 fps |
| API | OpenGL | OpenGL |
| OS | Windows | Windows |
| Tessellation | 5991.7 | 7000.8 |
Tech Specs
Specs & test comparisonGeneral Info
| Manufacturer | Nvidia | Nvidia |
| Build | Discrete | Discrete |
| Announced | January, 7, 2025 | March, 29, 2022 |
| Form Factor | PC use | PC use |
| Use case | Gaming | Gaming |
| Segment | High-end | High-end |
| Architecture | Blackwell 2.0 | Ampere |
| Codename | GB202-300-A1 | GA102 |
| Price | 1999 | 1999 |
GPU
| GPU Frequency | 2017 MHz | 1560 MHz |
| Boost GPU Frequency | 2407 MHz | 1860 MHz |
| Shader Units | 21760 | 10752 |
| TMUs | 680 | 336 |
| ROPs | 176 | 112 |
| Execution Units | 170 | 84 |
| Ray Tracing Cores | 170 | 84 |
| Tensor Cores | 680 | 336 |
| L1 cache | 128 KB/core | 128 KB/core |
| L2 cache | 88 KB/core | 6000 KB/core |
| Instructions per Clock | 2 IPC | 2 IPC |
Performance
| Pixel Fill Rate | 424 Gpix/sec | 208 Gpix/sec |
| Texture Fill Rate | 1637 GTex/s | 625 GTex/s |
| FLOPS | 104.8 TFLOPS | 40 TFLOPS |
Process
| Interface | PCIe 5.0 x16 | PCIe 4.0 x16 |
| TGP | 575 W | 450 W |
| Process | TSMC | Samsung |
| Process tech. | 4 nm | 8 nm |
| Transistor Size | 750 mm² | 628 mm² |
| Transistor count | 92.2 billion | 28 billion |
| Transistor density | 122.93 MTr/mm² | 44.59 MTr/mm² |
Memory
| Interface | GDDR7 | GDDR6X |
| Memory Frequency | 1750 MHz | 1313 MHz |
| Bus | 512 bit | 384 bit |
| ECC support | No | No |
| Effective Memory Speed | 28000 MB/s | 21000 MB/s |
| Memory Size | 32 GB | 24 GB |
Max Throughput
API
| DirectX | 12 | 12 |
| Vulkan | 1.4 | 1.3 |
| OpenGL | 4.6 | 4.6 |
| CUDA | 10.1 | 8.6 |
| DisplayPort | 2.1a | 1.4a |
| Ray Tracing | Yes | Yes |
| DLSS | DLSS 4 | DLSS 2 |
Results
After analyzing the data, we consider Nvidia GeForce RTX 5090 the clear winner in this comparison.
Survey
Total Votes: 0
