Product Compare
NVIDIA Rubin GPU vs NVIDIA Tesla P40
| GPU | GPU | |
|---|---|---|
| Graphics Card | ||
| Release Date | Jan 2026 | Sep 2016 |
| Generation | Server Rubin | Tesla Pascal |
| Type | Desktop | Professional |
| Bus Interface | PCIe 6.0 x16 | PCIe 3.0 x16 |
| Clock Speeds | ||
| Base Clock | 700 MHz | 1303 MHz |
| Boost Clock | 2267 MHz | 1531 MHz |
| Memory Clock | 2695 MHz | 1808 MHz |
| Memory | ||
| Memory Size | 288GB | 24GB |
| Memory Type | HBM4 | GDDR5 |
| Memory Bus | 16384bit | 384bit |
| Bandwidth | 22.1TB/s | 347.1GB/s |
| Render Config | ||
| SM Count | 224 | 30 |
| Shading Units | 28672 | 3840 |
| TMUs | 896 | 240 |
| ROPs | 24 | 96 |
| Tensor Cores | 896 | |
| L1 Cache | 256 KB (per SM) | 48 KB (per SM) |
| L2 Cache | 128 MB | 3 MB |
| Theoretical Performance | ||
| Pixel Rate | 54.41 GPixel/s | 147.0 GPixel/s |
| Texture Rate | 2031.2 GTexel/s | 367.4 GTexel/s |
| FP16 (half) | 260.0 TFLOPS | 183.7 GFLOPS |
| FP32 (float) | 130.0 TFLOPS | 11.76 TFLOPS |
| FP64 (double) | 32.50 TFLOPS | 367.4 GFLOPS |
| Graphics Processor | ||
| GPU Name | GR100 | GP102 |
| Architecture | Rubin | Pascal |
| Foundry | TSMC | TSMC |
| Process Size | 3 nm | 16 nm |
| Transistors | 336 billion | 11.8 billion |
| Die Size | 1456 mm² | 471 mm² |
| Board Design | ||
| TDP | 2300W | 250W |
| Suggested PSU | 2700 W | 600 W |
| Outputs | No outputs | No outputs |
| Power Connectors | 8-pin EPS | |
| Graphics Features | ||
| DirectX | N/A | 12 (12_1) |
| OpenGL | N/A | 4.6 |
| OpenCL | 3.0 | 3.0 |
| Vulkan | N/A | 1.3 |
| CUDA | 10.7 | 6.1 |
| Shader Model | N/A | 6.7 |