AMD Radeon Instinct MI100 vs NVIDIA A40 PCIe
What is the difference between AMD Radeon Instinct MI100 and NVIDIA A40 PCIe. Find out which graphics card has better performance.
Graphics Processor (GPU)
| Arcturus | GPU Name | GA102 |
| CDNA 1.0 | Architecture | Ampere |
| TSMC | Foundry | Samsung |
| 7 nm | Process Size | 8 nm |
| 50,000 million | Transistors | 28,300 million |
| 750 mm² | Die Size | 628 mm² |
Graphics Card
| Nov 16th, 2020 | Release Date | Oct 5th, 2020 |
| Radeon Instinct (MIx) | Family | Tesla (Axx) |
| Active | Production | Active |
| PCIe 4.0 x16 | Bus Interface | PCIe 4.0 x16 |
Memory
| 32 GB | Memory Size | 48 GB |
| HBM2 | Memory Type | GDDR6 |
| 4096 bit | Memory Bus | 384 bit |
| 1,229 GB/s | Bandwidth | 695.8 GB/s |
Performance
| 96.13 GPixel/s | Pixel fillrate | 194.9 GPixel/s |
| 721.0 GTexel/s | Texture fillrate | 584.6 GTexel/s |
| 184.6 TFLOPS (8:1) | FP16 (half) performance | 37.42 TFLOPS (1:1) |
| 23.07 TFLOPS | FP32 (float) performance | 37.42 TFLOPS |
| 11.54 TFLOPS (1:2) | FP64 (double) performance | 1,169 GFLOPS (1:32) |
Clock Speeds
| 1000 MHz | Base Clock | 1305 MHz |
| 1502 MHz | Boost Clock | 1740 MHz |
| 1200 MHz 2.4 Gbps effective | Memory Clock | 1812 MHz 14.5 Gbps effective |
Render Config
| 7680 | Shading Units | 10752 |
| 480 | TMUs | 336 |
| 64 | ROPs | 112 |
| 16 KB (per CU) | L1 Cache | 128 KB (per SM) |
| 8 MB | L2 Cache | 6 MB |
Board Design
| Dual-slot | Slot Width | Dual-slot |
| 267 mm 10.5 inches | Length | 267 mm 10.5 inches |
| 111 mm 4.4 inches | Width | 112 mm 4.4 inches |
| 300 W | Thermal design power (TDP) | 300 W |
| 700 W | Suggested PSU | 700 W |
| No outputs | Display Connectors | 3x DisplayPort |
| 2x 8-pin | Power Connectors | 8-pin EPS |
API support
| N/A | DirectX | 12 Ultimate (12_2) |
| N/A | OpenGL | 4.6 |
| 2.1 | OpenCL | 3.0 |
| N/A | Vulkan | 1.2 |
| N/A | Shader Model | 6.6 |