Product Compare
NVIDIA H100 PCIe 96 GB vs NVIDIA GeForce RTX 5090 DD
| GPU | GPU | |
|---|---|---|
| Graphics Card | ||
| Release Date | Mar 2023 | Unknown |
| Generation | Tesla Hopper | GeForce 50 |
| Type | AI GPU | Desktop |
| Bus Interface | PCIe 5.0 x16 | PCIe 5.0 x16 |
| Clock Speeds | ||
| Base Clock | 1665 MHz | 2017 MHz |
| Boost Clock | 1837 MHz | 2407 MHz |
| Memory Clock | 1313 MHz | 1750 MHz |
| Memory | ||
| Memory Size | 96GB | 24GB |
| Memory Type | HBM3 | GDDR7 |
| Memory Bus | 5120bit | 384bit |
| Bandwidth | 1681GB/s | 1.34TB/s |
| Render Config | ||
| SM Count | 132 | 170 |
| Shading Units | 16896 | 21760 |
| TMUs | 528 | 680 |
| ROPs | 24 | 176 |
| Tensor Cores | 528 | 680 |
| L1 Cache | 256 KB (per SM) | 128 KB (per SM) |
| L2 Cache | 50 MB | 96 MB |
| RT Cores | 170 | |
| Theoretical Performance | ||
| Pixel Rate | 44.09 GPixel/s | 423.6 GPixel/s |
| Texture Rate | 969.9 GTexel/s | 1636.8 GTexel/s |
| FP16 (half) | 248.3 TFLOPS | 104.8 TFLOPS |
| FP32 (float) | 62.08 TFLOPS | 104.8 TFLOPS |
| FP64 (double) | 31.04 TFLOPS | 1.637 TFLOPS |
| Graphics Processor | ||
| GPU Name | GH100 | GB202 |
| Architecture | Hopper | Blackwell 2.0 |
| Foundry | TSMC | TSMC |
| Process Size | 4 nm | 5 nm |
| Transistors | 80 billion | 92.2 billion |
| Die Size | 814 mm² | 750 mm² |
| GPU Variant | GB202-240-K1-A1 | |
| Board Design | ||
| TDP | 700W | 575W |
| Suggested PSU | 1100 W | 950 W |
| Outputs | No outputs | 1x HDMI 2.1b3x DisplayPort 2.1b |
| Power Connectors | 8-pin EPS | 1x 16-pin |
| Graphics Features | ||
| DirectX | N/A | 12 Ultimate (12_2) |
| OpenGL | N/A | 4.6 |
| OpenCL | 3.0 | 3.0 |
| Vulkan | N/A | 1.4 |
| CUDA | 9.0 | 12.0 |
| Shader Model | N/A | 6.8 |