Product Compare
NVIDIA CMP 40HX vs NVIDIA L40
| GPU | GPU | |
|---|---|---|
| Graphics Card | ||
| Release Date | Feb 2021 | Oct 2022 |
| Generation | Mining GPUs | Tesla Ada |
| Type | Desktop | Professional |
| Bus Interface | PCIe 1.0 x4 | PCIe 4.0 x16 |
| Clock Speeds | ||
| Base Clock | 1470 MHz | 735 MHz |
| Boost Clock | 1650 MHz | 2490 MHz |
| Memory Clock | 1750 MHz | 2250 MHz |
| Memory | ||
| Memory Size | 8GB | 48GB |
| Memory Type | GDDR6 | GDDR6 |
| Memory Bus | 256bit | 384bit |
| Bandwidth | 448.0GB/s | 864.0GB/s |
| Render Config | ||
| SM Count | 36 | 142 |
| Shading Units | 2304 | 18176 |
| TMUs | 144 | 568 |
| ROPs | 64 | 192 |
| Tensor Cores | 288 | 568 |
| RT Cores | 36 | 142 |
| L1 Cache | 64 KB (per SM) | 128 KB (per SM) |
| L2 Cache | 4 MB | 96 MB |
| Theoretical Performance | ||
| Pixel Rate | 105.6 GPixel/s | 478.1 GPixel/s |
| Texture Rate | 237.6 GTexel/s | 1414 GTexel/s |
| FP16 (half) | 15.21 TFLOPS | 90.52 TFLOPS |
| FP32 (float) | 7.603 TFLOPS | 90.52 TFLOPS |
| FP64 (double) | 237.6 GFLOPS | 1414 GFLOPS |
| Graphics Processor | ||
| GPU Name | TU106 | AD102 |
| GPU Variant | TU106-100-A1 | AD102-895-A1 |
| Architecture | Turing | Ada Lovelace |
| Foundry | TSMC | TSMC |
| Process Size | 12 nm | 4 nm |
| Transistors | 10.8 billion | 76.3 billion |
| Die Size | 445 mm² | 608 mm² |
| Board Design | ||
| TDP | 185W | 300W |
| Suggested PSU | 450 W | 700 W |
| Outputs | No outputs | 4x DisplayPort 1.4a |
| Power Connectors | 1x 8-pin | 1x 16-pin |
| Graphics Features | ||
| DirectX | 12 Ultimate (12_2) | 12 Ultimate (12_2) |
| OpenGL | 4.6 | 4.6 |
| OpenCL | 3.0 | 3.0 |
| Vulkan | 1.3 | 1.3 |
| CUDA | 7.5 | 8.9 |
| Shader Model | 6.8 | 6.6 |
| Benchmarks | ||
| Blender | 1174 | 3901 |