AMD Radeon RX 560 896SP vs NVIDIA GRID M6-8Q
Processeur graphique
| Nom du GPU | Polaris 21 | GM204 |
|---|---|---|
| GPU variant | Polaris 21 XL | GM204-995-A1 |
| Architecture | GCN 4.0 | Maxwell 2.0 |
| Fonderie | GlobalFoundries | TSMC |
| Taille du processus | 14 nm | 28 nm |
| Transistors | 3,000 million | 5,200 million |
| Taille des matrices | 123 mm² | 398 mm² |
Carte graphique
| Date de publication | Jul 4th, 2017 | Aug 30th, 2015 |
|---|---|---|
| Génération | Polaris | GRID |
| Production | Active | End-of-life |
| Bus interface | PCIe 3.0 x8 | PCIe 3.0 x16 |
| Prédécesseur | Arctic Islands | — |
| Successeur | Vega | — |
Vitesse clock
| Base clock | 1090 MHz | — |
|---|---|---|
| Boost clock | 1175 MHz | — |
| Mémoire clock | 1750 MHz 7 Gbps effective | 1253 MHz 5 Gbps effective |
| GPU clock | — | 722 MHz |
Mémoire
| Taille de la mémoire | 4 GB | 8 GB |
|---|---|---|
| Type de mémoire | GDDR5 | GDDR5 |
| Bus mémoire | 128 bit | 256 bit |
| Bande passante | 112.0 GB/s | 160.4 GB/s |
Render Config
| Shading units | 896 | 1536 |
|---|---|---|
| TMUs | 56 | 96 |
| ROPs | 16 | 64 |
| Compute units | 14 | — |
| Cache L1 | 16 KB (per CU) | 48 KB (per SMM) |
| Cache L2 | 1024 KB | 2 MB |
| Comptage SMM | — | 12 |
Performance théorique
| Taux de pixel | 18.80 GPixel/s | 46.21 GPixel/s |
|---|---|---|
| Taux de texture | 65.80 GTexel/s | 69.31 GTexel/s |
| FP16 (half) performance | 2.106 TFLOPS (1:1) | — |
| FP32 (float) performance | 2.106 TFLOPS | 2.218 TFLOPS |
| FP64 (double) performance | 131.6 GFLOPS (1:16) | 69.31 GFLOPS (1:32) |
Board Design
| Taille Slot | Dual-slot | MXM Module |
|---|---|---|
| Longueur | 170 mm 6.7 inches | — |
| TDP | 45 W | 100 W |
| Suggestions pour le PSU | 200 W | 300 W |
| Sorties | 1x DVI1x HDMI1x DisplayPort | No outputs |
| Connecteurs de puissance | None | None |
| Board number | C981-21 C994-71 | — |
Caractéristiques graphiques
| DirectX | 12 (12_0) | 12 (12_1) |
|---|---|---|
| OpenGL | 4.6 | 4.6 |
| OpenCL | 2.1 | 3.0 |
| Vulkan | 1.2 | 1.1 |
| Shader model | 6.4 | 6.4 |
| CUDA | — | 5.2 |
