Product Compare
NVIDIA H800 SXM5 vs NVIDIA Tesla T40 24 GB
| GPU | GPU | |
|---|---|---|
| Graphics Card | ||
| Release Date | Mar 2023 | Unknown |
| Generation | Server Hopper | Tesla Turing |
| Type | AI GPU | Professional |
| Bus Interface | PCIe 5.0 x16 | PCIe 3.0 x16 |
| Clock Speeds | ||
| Base Clock | 1095 MHz | 1305 MHz |
| Boost Clock | 1755 MHz | 1560 MHz |
| Memory Clock | 1313 MHz | 1625 MHz |
| Memory | ||
| Memory Size | 80GB | 24GB |
| Memory Type | HBM3 | GDDR6 |
| Memory Bus | 5120bit | 384bit |
| Bandwidth | 3.36TB/s | 624.0GB/s |
| Render Config | ||
| SM Count | 132 | 72 |
| Shading Units | 16896 | 4608 |
| TMUs | 528 | 288 |
| ROPs | 24 | 96 |
| Tensor Cores | 528 | 576 |
| L1 Cache | 256 KB (per SM) | 64 KB (per SM) |
| L2 Cache | 50 MB | 6 MB |
| RT Cores | 72 | |
| Theoretical Performance | ||
| Pixel Rate | 42.12 GPixel/s | 149.8 GPixel/s |
| Texture Rate | 926.6 GTexel/s | 449.3 GTexel/s |
| FP16 (half) | 237.2 TFLOPS | 28.75 TFLOPS |
| FP32 (float) | 59.30 TFLOPS | 14.38 TFLOPS |
| FP64 (double) | 29.65 TFLOPS | 449.3 GFLOPS |
| Graphics Processor | ||
| GPU Name | GH100 | TU102 |
| Architecture | Hopper | Turing |
| Foundry | TSMC | TSMC |
| Process Size | 5 nm | 12 nm |
| Transistors | 80 billion | 18.6 billion |
| Die Size | 814 mm² | 754 mm² |
| Board Design | ||
| TDP | 700W | 250W |
| Suggested PSU | 1100 W | 600 W |
| Outputs | No outputs | No outputs |
| Power Connectors | 8-pin EPS | 1x 8-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 | 7.5 |
| Shader Model | N/A | 6.8 |