AMD Radeon Pro Vega 20 vs NVIDIA GeForce RTX 2070 Max-Q Refresh
Processeur graphique
| Nom du GPU | Vega 12 | TU106B |
|---|---|---|
| GPU variant | Vega 12 XTA | — |
| Architecture | GCN 5.0 | Turing |
| Fonderie | GlobalFoundries | TSMC |
| Taille du processus | 14 nm | 12 nm |
| Transistors | unknown | 10,800 million |
| Taille des matrices | unknown | 445 mm² |
Graphiques mobiles
| Date de publication | Nov 14th, 2018 | Mar 4th, 2020 |
|---|---|---|
| Génération | Radeon Pro Mac (Vega Series) | GeForce 20 Mobile |
| Production | Active | Active |
| Bus interface | PCIe 3.0 x16 | PCIe 3.0 x16 |
| Prédécesseur | — | GeForce 10 Mobile |
| Successeur | — | GeForce 30 Mobile |
Vitesse clock
| Base clock | 815 MHz | 900 MHz |
|---|---|---|
| Boost clock | 1283 MHz | 1125 MHz |
| Mémoire clock | 740 MHz 1480 Mbps effective | 1375 MHz 11 Gbps effective |
Mémoire
| Taille de la mémoire | 4 GB | 8 GB |
|---|---|---|
| Type de mémoire | HBM2 | GDDR6 |
| Bus mémoire | 1024 bit | 256 bit |
| Bande passante | 189.4 GB/s | 352.0 GB/s |
Render Config
| Shading units | 1280 | 2304 |
|---|---|---|
| TMUs | 80 | 144 |
| ROPs | 32 | 64 |
| Compute units | 20 | — |
| Cache L1 | 16 KB (per CU) | 64 KB (per SM) |
| Cache L2 | 1024 KB | 4 MB |
| Comptage SM | — | 36 |
| Tensor cores | — | 288 |
| RT cores | — | 36 |
Performance théorique
| Taux de pixel | 41.06 GPixel/s | 72.00 GPixel/s |
|---|---|---|
| Taux de texture | 102.6 GTexel/s | 162.0 GTexel/s |
| FP16 (half) performance | 6.569 TFLOPS (2:1) | 10.37 TFLOPS (2:1) |
| FP32 (float) performance | 3.284 TFLOPS | 5.184 TFLOPS |
| FP64 (double) performance | 205.3 GFLOPS (1:16) | 162.0 GFLOPS (1:32) |
Board Design
| Taille Slot | IGP | MXM Module |
|---|---|---|
| TDP | 100 W | 115 W |
| Sorties | No outputs | No outputs |
| Connecteurs de puissance | — | None |
| Board number | — | E4914 SKU 31 |
Caractéristiques graphiques
| DirectX | 12 (12_1) | 12 Ultimate (12_2) |
|---|---|---|
| OpenGL | 4.6 | 4.6 |
| OpenCL | 2.1 | 3.0 |
| Vulkan | 1.2 | 1.2 |
| Shader model | 6.3 | 6.6 |
| CUDA | — | 7.5 |
