Product Compare
NVIDIA L40S vs NVIDIA GeForce RTX 3090 Ti
| GPU | GPU | |
|---|---|---|
| Graphics Card | ||
| Release Date | Oct 2022 | Jan 2022 |
| Generation | Server Ada | GeForce 30 |
| Type | Professional | Desktop |
| Bus Interface | PCIe 4.0 x16 | PCIe 4.0 x16 |
| Clock Speeds | ||
| Base Clock | 1065 MHz | 1560 MHz |
| Boost Clock | 2520 MHz | 1860 MHz |
| Memory Clock | 2250 MHz | 1313 MHz |
| Memory | ||
| Memory Size | 48GB | 24GB |
| Memory Type | GDDR6 | GDDR6X |
| Memory Bus | 384bit | 384bit |
| Bandwidth | 864.0GB/s | 1.01TB/s |
| Render Config | ||
| SM Count | 142 | 84 |
| Shading Units | 18176 | 10752 |
| TMUs | 568 | 336 |
| ROPs | 192 | 112 |
| Tensor Cores | 568 | 336 |
| RT Cores | 142 | 84 |
| L1 Cache | 128 KB (per SM) | 128 KB (per SM) |
| L2 Cache | 48 MB | 6 MB |
| Theoretical Performance | ||
| Pixel Rate | 483.8 GPixel/s | 208.3 GPixel/s |
| Texture Rate | 1431.4 GTexel/s | 625.0 GTexel/s |
| FP16 (half) | 91.61 TFLOPS | 40.00 TFLOPS |
| FP32 (float) | 91.61 TFLOPS | 40.00 TFLOPS |
| FP64 (double) | 1431.4 GFLOPS | 625.0 GFLOPS |
| Graphics Processor | ||
| GPU Name | AD102 | GA102 |
| GPU Variant | AD102-896-A1 | GA102-350-A1 |
| Architecture | Ada Lovelace | Ampere |
| Foundry | TSMC | Samsung |
| Process Size | 5 nm | 8 nm |
| Transistors | 76.3 billion | 28.3 billion |
| Die Size | 609 mm² | 628 mm² |
| Board Design | ||
| TDP | 350W | 450W |
| Suggested PSU | 750 W | 850 W |
| Outputs | 4x DisplayPort 1.4a | 1x HDMI 2.13x DisplayPort 1.4a |
| Power Connectors | 1x 16-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 | 1.4 | 1.4 |
| CUDA | 8.9 | 8.6 |
| Shader Model | 6.9 | 6.8 |
| Benchmarks | ||
| Blender | 9239 | 6217 |