Product Compare
NVIDIA H100 PCIe vs NVIDIA GeForce RTX 3070 Ti GA102
| GPU | GPU | |
|---|---|---|
| Graphics Card | ||
| Release Date | Mar 2022 | Oct 2022 |
| Generation | Tesla Hopper | GeForce 30 |
| Type | AI GPU | Desktop |
| Bus Interface | PCIe 5.0 x16 | PCIe 4.0 x16 |
| Clock Speeds | ||
| Base Clock | 1095 MHz | 1575 MHz |
| Boost Clock | 1755 MHz | 1770 MHz |
| Memory Clock | 1593 MHz | 1188 MHz |
| Memory | ||
| Memory Size | 80GB | 8GB |
| Memory Type | HBM2e | GDDR6X |
| Memory Bus | 5120bit | 256bit |
| Bandwidth | 2039GB/s | 608.3GB/s |
| Render Config | ||
| SM Count | 114 | 48 |
| Shading Units | 14592 | 6144 |
| TMUs | 456 | 192 |
| ROPs | 24 | 96 |
| Tensor Cores | 456 | 192 |
| L1 Cache | 256 KB (per SM) | 128 KB (per SM) |
| L2 Cache | 50 MB | 4 MB |
| RT Cores | 48 | |
| Theoretical Performance | ||
| Pixel Rate | 42.12 GPixel/s | 169.9 GPixel/s |
| Texture Rate | 800.3 GTexel/s | 339.8 GTexel/s |
| FP16 (half) | 204.9 TFLOPS | 21.75 TFLOPS |
| FP32 (float) | 51.22 TFLOPS | 21.75 TFLOPS |
| FP64 (double) | 25.61 TFLOPS | 339.8 GFLOPS |
| Graphics Processor | ||
| GPU Name | GH100 | GA102 |
| Architecture | Hopper | Ampere |
| Foundry | TSMC | Samsung |
| Process Size | 4 nm | 8 nm |
| Transistors | 80 billion | 28.3 billion |
| Die Size | 814 mm² | 628 mm² |
| GPU Variant | GA102-150-A1 | |
| Board Design | ||
| TDP | 350W | 290W |
| Suggested PSU | 750 W | 600 W |
| Outputs | No outputs | 1x HDMI 2.13x DisplayPort 1.4a |
| Power Connectors | 1x 16-pin | 1x 12-pin |
| Graphics Features | ||
| DirectX | N/A | 12 Ultimate (12_2) |
| OpenGL | N/A | 4.6 |
| OpenCL | 3.0 | 3.0 |
| Vulkan | N/A | 1.3 |
| CUDA | 9.0 | 8.6 |
| Shader Model | N/A | 6.6 |