Product Compare
NVIDIA Tesla K40st vs AMD Radeon RX 560 896SP
| GPU | GPU | |
|---|---|---|
| Graphics Card | ||
| Release Date | Nov 2013 | Jul 2017 |
| Generation | Tesla | Polaris |
| Type | Professional | Desktop |
| Bus Interface | PCIe 3.0 x16 | PCIe 3.0 x8 |
| Clock Speeds | ||
| Memory Clock | 1502 MHz | 1750 MHz |
| Base Clock | 1090 MHz | |
| Boost Clock | 1175 MHz | |
| Memory | ||
| Memory Size | 12GB | 4GB |
| Memory Type | GDDR5 | GDDR5 |
| Memory Bus | 384bit | 128bit |
| Bandwidth | 288.4GB/s | 112.0GB/s |
| Render Config | ||
| Shading Units | 2880 | 896 |
| TMUs | 240 | 56 |
| ROPs | 48 | 16 |
| L1 Cache | 16 KB (per SMX) | 16 KB (per CU) |
| L2 Cache | 1536 KB | 1024 KB |
| Compute Units | 14 | |
| Theoretical Performance | ||
| Pixel Rate | 34.50 GPixel/s | 18.80 GPixel/s |
| Texture Rate | 138.0 GTexel/s | 65.80 GTexel/s |
| FP32 (float) | 3.312 TFLOPS | 2.106 TFLOPS |
| FP64 (double) | 1104 GFLOPS | 131.6 GFLOPS |
| FP16 (half) | 2.106 TFLOPS | |
| Graphics Processor | ||
| GPU Name | GK110B | Polaris 21 |
| Architecture | Kepler | GCN 4.0 |
| Foundry | TSMC | GlobalFoundries |
| Process Size | 28 nm | 14 nm |
| Transistors | 7.08 billion | 3 billion |
| Die Size | 561 mm² | 123 mm² |
| GPU Variant | Polaris 21 XL | |
| Board Design | ||
| TDP | 245W | 45W |
| Suggested PSU | 550 W | 200 W |
| Outputs | No outputs | 1x DVI1x HDMI 2.0b1x DisplayPort 1.4a |
| Power Connectors | None | |
| Graphics Features | ||
| DirectX | 12 (11_1) | 12 (12_0) |
| OpenGL | 4.6 | 4.6 |
| OpenCL | 3.0 | 2.1 |
| Vulkan | 1.1 | 1.2 |
| CUDA | 3.5 | |
| Shader Model | 5.1 | 6.4 |