NVIDIA A100 PCIe vs NVIDIA A40 PCIe
What is the difference between NVIDIA A100 PCIe and NVIDIA A40 PCIe. Find out which graphics card has better performance.
Graphics Processor (GPU)
| GA100 | GPU Name | GA102 |
| Ampere | Architecture | Ampere |
| TSMC | Foundry | Samsung |
| 7 nm | Process Size | 8 nm |
| 54,200 million | Transistors | 28,300 million |
| 826 mm² | Die Size | 628 mm² |
Graphics Card
| Jun 22nd, 2020 | Release Date | Oct 5th, 2020 |
| Tesla (Axx) | Family | Tesla (Axx) |
| Active | Production | Active |
| PCIe 4.0 x16 | Bus Interface | PCIe 4.0 x16 |
Memory
| 40 GB | Memory Size | 48 GB |
| HBM2e | Memory Type | GDDR6 |
| 5120 bit | Memory Bus | 384 bit |
| 1,555 GB/s | Bandwidth | 695.8 GB/s |
Performance
| 225.6 GPixel/s | Pixel fillrate | 194.9 GPixel/s |
| 609.1 GTexel/s | Texture fillrate | 584.6 GTexel/s |
| 77.97 TFLOPS (4:1) | FP16 (half) performance | 37.42 TFLOPS (1:1) |
| 19.49 TFLOPS | FP32 (float) performance | 37.42 TFLOPS |
| 9.746 TFLOPS (1:2) | FP64 (double) performance | 1,169 GFLOPS (1:32) |
Clock Speeds
| 765 MHz | Base Clock | 1305 MHz |
| 1410 MHz | Boost Clock | 1740 MHz |
| 1215 MHz 2.4 Gbps effective | Memory Clock | 1812 MHz 14.5 Gbps effective |
Render Config
| 6912 | Shading Units | 10752 |
| 432 | TMUs | 336 |
| 160 | ROPs | 112 |
| 192 KB (per SM) | L1 Cache | 128 KB (per SM) |
| 40 MB | L2 Cache | 6 MB |
| 108 | SM Count | 84 |
| 432 | Tensor Cores | 336 |
Board Design
| Dual-slot | Slot Width | Dual-slot |
| 267 mm 10.5 inches | Length | 267 mm 10.5 inches |
| 250 W | Thermal design power (TDP) | 300 W |
| 600 W | Suggested PSU | 700 W |
| No outputs | Display Connectors | 3x DisplayPort |
| 8-pin EPS | Power Connectors | 8-pin EPS |
API support
| N/A | DirectX | 12 Ultimate (12_2) |
| N/A | OpenGL | 4.6 |
| 3.0 | OpenCL | 3.0 |
| N/A | Vulkan | 1.2 |
| N/A | Shader Model | 6.6 |
| 8.0 | CUDA | 8.6 |