Product Compare
NVIDIA RTX PRO 4000 Blackwell vs NVIDIA Tesla K20m
| GPU | GPU | |
|---|---|---|
| Graphics Card | ||
| Release Date | Mar 2025 | Jan 2013 |
| Generation | Blackwell PRO W | Tesla |
| Type | Desktop | Professional |
| Bus Interface | PCIe 5.0 x16 | PCIe 2.0 x16 |
| Clock Speeds | ||
| Base Clock | 1230 MHz | |
| Boost Clock | 2055 MHz | |
| Memory Clock | 1750 MHz | 1300 MHz |
| Memory | ||
| Memory Size | 24GB | 5GB |
| Memory Type | GDDR7 | GDDR5 |
| Memory Bus | 192bit | 320bit |
| Bandwidth | 672.0GB/s | 208.0GB/s |
| Render Config | ||
| SM Count | 70 | |
| Shading Units | 8960 | 2496 |
| TMUs | 280 | 208 |
| ROPs | 96 | 40 |
| Tensor Cores | 280 | |
| RT Cores | 70 | |
| L1 Cache | 128 KB (per SM) | 16 KB (per SMX) |
| L2 Cache | 48 MB | 1280 KB |
| Theoretical Performance | ||
| Pixel Rate | 197.3 GPixel/s | 36.71 GPixel/s |
| Texture Rate | 575.4 GTexel/s | 146.8 GTexel/s |
| FP16 (half) | 36.83 TFLOPS | |
| FP32 (float) | 36.83 TFLOPS | 3.524 TFLOPS |
| FP64 (double) | 575.4 GFLOPS | 1175 GFLOPS |
| Graphics Processor | ||
| GPU Name | GB203 | GK110 |
| Architecture | Blackwell 2.0 | Kepler |
| Foundry | TSMC | TSMC |
| Process Size | 5 nm | 28 nm |
| Transistors | 45.6 billion | 7.08 billion |
| Die Size | 378 mm² | 561 mm² |
| Board Design | ||
| TDP | 140W | 225W |
| Suggested PSU | 300 W | 550 W |
| Outputs | 4x DisplayPort 2.1b | No outputs |
| Power Connectors | 1x 16-pin | 1x 6-pin + 1x 8-pin |
| Graphics Features | ||
| DirectX | 12 Ultimate (12_2) | 12 (11_0) |
| OpenGL | 4.6 | 4.6 |
| OpenCL | 3.0 | 3.0 |
| Vulkan | 1.4 | 1.1 |
| CUDA | 12.0 | 3.5 |
| Shader Model | 6.8 | 5.1 |