AMD Radeon R9 M395X Mac Edition vs NVIDIA GeForce RTX 2070 Max-Q Refresh
Graphics Processor
| GPU model | Amethyst | TU106B |
|---|---|---|
| GPU variant | Amethyst XT | — |
| Architecture | GCN 3.0 | Turing |
| Foundry | TSMC | TSMC |
| Manufacturing process | 28 nm | 12 nm |
| Number of transistors | 5,000 million | 10,800 million |
| Die size | 366 mm² | 445 mm² |
Mobile graphics
| Release date | May 5th, 2015 | Mar 4th, 2020 |
|---|---|---|
| Generation | Crystal System (R9 M300) | GeForce 20 Mobile |
| Production | End-of-life | Active |
| Interface | MXM-B (3.0) | PCIe 3.0 x16 |
| Predecessor | Solar System | GeForce 10 Mobile |
| Successor | Mobility Radeon | GeForce 30 Mobile |
Clocks
| GPU clock | 909 MHz | — |
|---|---|---|
| Memory clock | 1365 MHz 5.5 Gbps effective | 1375 MHz 11 Gbps effective |
| Base clock | — | 900 MHz |
| Boost clock | — | 1125 MHz |
Memory Configuration
| Memory size | 4 GB | 8 GB |
|---|---|---|
| Memory type | GDDR5 | GDDR6 |
| Memory bus width | 256 bit | 256 bit |
| Bandwidth | 174.7 GB/s | 352.0 GB/s |
Render Configuration
| Shading units | 2048 | 2304 |
|---|---|---|
| TMUs | 128 | 144 |
| ROPs | 32 | 64 |
| Compute units | 32 | — |
| Cache L1 | 16 KB (per CU) | 64 KB (per SM) |
| Cache L2 | 512 KB | 4 MB |
| SM count | — | 36 |
| Tensor cores | — | 288 |
| RT cores | — | 36 |
Performance
| Pixel rate | 29.09 GPixel/s | 72.00 GPixel/s |
|---|---|---|
| Texture rate | 116.4 GTexel/s | 162.0 GTexel/s |
| FP16 (half) performance | 3.723 TFLOPS (1:1) | 10.37 TFLOPS (2:1) |
| FP32 (float) performance | 3.723 TFLOPS | 5.184 TFLOPS |
| FP64 (double) performance | 232.7 GFLOPS (1:16) | 162.0 GFLOPS (1:32) |
Dimensions & Outputs
| Slot width | MXM Module | MXM Module |
|---|---|---|
| TDP | 250 W | 115 W |
| Outputs | No outputs | No outputs |
| Power connectors | None | None |
| Board number | — | E4914 SKU 31 |
API Support & Features
| DirectX | 12 (12_0) | 12 Ultimate (12_2) |
|---|---|---|
| OpenGL | 4.6 | 4.6 |
| OpenCL | 2.0 | 3.0 |
| Vulkan | 1.2 | 1.2 |
| Shader model | 6.3 | 6.6 |
| CUDA | — | 7.5 |
