AMD Radeon E9173 PCIe vs NVIDIA GeForce GTX 670M
Processeur graphique
| Nom du GPU | Lexa | GF114 |
|---|---|---|
| GPU variant | E9170 | N13E-GS1-LP-A1 |
| Architecture | GCN 4.0 | Fermi 2.0 |
| Fonderie | GlobalFoundries | TSMC |
| Taille du processus | 14 nm | 40 nm |
| Transistors | 2,200 million | 1,950 million |
| Taille des matrices | 103 mm² | 332 mm² |
Graphiques mobiles
| Date de publication | Oct 3rd, 2017 | Mar 22nd, 2012 |
|---|---|---|
| Génération | Embedded (9000) | GeForce 600M |
| Production | End-of-life | End-of-life |
| Bus interface | PCIe 3.0 x8 | MXM-B (3.0) |
| Prédécesseur | — | GeForce 500M |
| Successeur | — | GeForce 700M |
Vitesse clock
| Base clock | 1124 MHz | — |
|---|---|---|
| Boost clock | 1219 MHz | — |
| Mémoire clock | 1500 MHz 6 Gbps effective | 750 MHz 3 Gbps effective |
| GPU clock | — | 598 MHz |
| Shader clock | — | 1196 MHz |
Mémoire
| Taille de la mémoire | 2 GB | 1536 MB |
|---|---|---|
| Type de mémoire | GDDR5 | GDDR5 |
| Bus mémoire | 64 bit | 192 bit |
| Bande passante | 48.00 GB/s | 72.00 GB/s |
Render Config
| Shading units | 512 | 336 |
|---|---|---|
| TMUs | 32 | 56 |
| ROPs | 16 | 24 |
| Compute units | 8 | — |
| Cache L1 | 16 KB (per CU) | 64 KB (per SM) |
| Cache L2 | 256 KB | 384 KB |
| Comptage SM | — | 7 |
Performance théorique
| Taux de pixel | 19.50 GPixel/s | 8.372 GPixel/s |
|---|---|---|
| Taux de texture | 39.01 GTexel/s | 33.49 GTexel/s |
| FP16 (half) performance | 1,248 GFLOPS (1:1) | — |
| FP32 (float) performance | 1,248 GFLOPS | 803.7 GFLOPS |
| FP64 (double) performance | 78.02 GFLOPS (1:16) | 66.98 GFLOPS (1:12) |
Board Design
| Taille Slot | Single-slot | MXM Module |
|---|---|---|
| TDP | 35 W | 75 W |
| Sorties | 1x DisplayPort2x mini-DisplayPort | No outputs |
| Connecteurs de puissance | None | — |
| Numéro de pièce | 102-D09126 | — |
| Board number | 109-D09187 | — |
Caractéristiques graphiques
| DirectX | 12 (12_0) | 12 (11_0) |
|---|---|---|
| OpenGL | 4.6 | 4.6 |
| OpenCL | 2.1 | 1.1 |
| Vulkan | 1.2 | — |
| Shader model | 6.4 | 5.1 |
| CUDA | — | 2.1 |
