GeForce RTX 5090
Score
Intel Arc B390 | Gaming | |
| Workstation | |
| AI/ML | |
| Energy Efficiency | |
| Quick Comparison Final Score | |
GeForce RTX 5090
Summary
Intel Arc B390| GB202-300 | Graphics Processor | Panther Lake |
| 21760 | Cores | 1536 |
| 680 / 176 | TMUs / ROPs | 48 / 24 |
GeForce RTX 5090
Blender
Intel Arc B390| 15010.34 | Blender GPU | 1268.79 |
GeForce RTX 5090
Games
Intel Arc B390| 244 | Forza Horizon 5 | 38 |
| 323 | The Witcher 3 | 55 |
| 265 | Counter-Strike 2 | 60 |
| 202 | Far Cry 6 | 49 |
| 207 | Hogwarts Legacy | 21 |
| 327 | Call of Duty | 32 |
| 179 | Ghost of Tsushima | 25 |
| 229 | Cyberpunk 2077 | 33 |
| 299 | Tomb Raider | 52 |
| 253 | Average FPS | 41 |
GeForce RTX 5090
Average FPS by Resolution
Intel Arc B390| 294 | 1080p High | 52 |
| 253 | 1080p Ultra | 41 |
| 207 | 1440p Ultra | 29 |
| 124 | 4K Ultra | 14 |
| Low | Margin of Error | Low |
GeForce RTX 5090
GeekBench 6 OpenCL
Intel Arc B390| 419309 | GB6 Compute Score | 64546 |
| 333 img/sec | Background Blur | — |
| 298.5 img/sec | Face Detection | — |
| 21.7 Gpixels/sec | Horizon Detection | — |
| 33.7 Gpixels/sec | Edge Detection | — |
| 34.5 Gpixels/sec | Gaussian Blur | — |
| 2.53 Gpixels/sec | Feature Matching | — |
| 2850 Gpixels/sec | Stereo Matching | — |
| 59146.9 FPS | Particle Physics | — |
| OpenCL | API | — |
GeForce RTX 5090
Passmark Graphics
Intel Arc B390| 38955 | G3D Mark Score | 9029 |
| 1413 | G2D Mark | 786 |
| 341 FPS | DirectX 11 | 70 FPS |
| 178 FPS | DirectX 12 | 50 FPS |
| 24632 Ops/s | GPU Compute | 3702 Ops/s |
GeForce RTX 5090
GeekBench 6 ML
Intel Arc B390| 114685 | Face Detection (HP) | 25029 |
| 344729 | Pose Estimation (Q) | 18852 |
| 4464 | Text Classification (SP) | 2115 |
| 8184 | Machine Translation (HP) | 4266 |
| 27802 | Object Detection (SP) | 7449 |
| 85572 | Depth Estimation (Q) | 22207 |
| 687080 | Style Transfer (SP) | 95039 |
| ONNX | Framework | ONNX |
| DirectML | Backend | DirectML |
| 53861 | GB6 ML Single Precision | 11871 |
| 81080 | GB6 ML Half Precision | 20120 |
| 38938 | GB6 ML Quantized | 7879 |
| 18948 | Image Classification (SP) | 5487 |
| 55109 | Image Segmentation (HP) | 16864 |
| 58478 | Image Super Resolution (Q) | 8788 |
GeForce RTX 5090
Raw Performance
Intel Arc B390| 424 GPixel/s | Pixel Fill Rate | 60 GPixel/s |
| 1637 GTexel/s | Texture Fill Rate | 120 GTexel/s |
| 104.8 TFLOPS | FLOPS (FP32) | 7.7 TFLOPS |
GeForce RTX 5090
Memory
Intel Arc B390| GDDR7 | Memory Type | System Shared |
| 32 GB | Memory Size | — |
| 1750 MHz | Memory Clock | — |
| 28000 Mbps | Effective Memory Speed | — |
| 512-bit | Bus | — |
| No | ECC | No |
| 1792 GB/s | Memory Bandwidth | — |
GeForce RTX 5090
Physical
Intel Arc B390| PCIe 5.0 x16 | Interface | PCIe 5.0 |
| 575 W | TGP | — |
| TSMC | Manufacturing | TSMC |
| 4 nm | Fabrication Process | 3 nm |
| 750 mm² | Die Size | — |
| 92.2 billion | Transistor Count | — |
| 122.93 MTr/mm² | Transistor Density | — |
GeForce RTX 5090
API
Intel Arc B390| 12.2 | DirectX | 12.1 |
| 1.4 | Vulkan | 1.4 |
| 4.6 | OpenGL | 4.6 |
| 3 | OpenCL | 3 |
| 12 | CUDA | — |
| Yes | Ray Tracing | Yes |
| DLSS 4 | DLSS | No |
| 2.1b | DisplayPort | 2.1 |
GeForce RTX 5090
General
Intel Arc B390| Nvidia | Vendor | Intel |
| Discrete | Build | Integrated |
| January 7, 2025 | Released | January 5, 2026 |
| Desktop | Case | Laptop |
| Gaming | Purpose | Gaming |
| High-end | Segment | Mid-range |
| Blackwell 2.0 | Architecture | Celestial |
| GB202-300 | GPU Codename | Panther Lake |
| -Intel Core Ultra 9 285Kor above | Recommended CPU | -Intel Core Ultra X9 388Hor above |
GeForce RTX 5090
3D Mark
Intel Arc B390| 52638 | Steel Nomad Lite Score | 4242 |
| 46941 | Time Spy | 6595 |
| 235224 | Solar Bay | 26399 |
| 38113 | Port Royal | 3886 |
| 87433 | Fire Strike | 16770 |
| 108736 | Wild Life Extreme | 11707 |
| 207058 | Night Raid | 68498 |
GeForce RTX 5090
Graphics Processing Unit
Intel Arc B390| 2017 MHz | Base Clock | 300 MHz |
| 2407 MHz | Boost Clock | 2500 MHz |
| 21760 | Shading Units | 1536 |
| 680 | Texture Mapping Units (TMUs) | 48 |
| 176 | Render Output Units (ROPs) | 24 |
| 170 | Compute Units (Pipelines) | 12 |
| 680 | Tensor Cores | — |
| 170 | Ray-tracing Cores | — |
| 128KB per cluster | L1 Cache | 64KB per cluster |
| 96MB shared | L2 Cache | 16MB shared |
| 2 IPC | Instructions Per Cycle | 2 IPC |
