AMD Radeon Instinct MI8 vs NVIDIA Quadro P1000
Processeur graphique
| Nom du GPU | Fiji | GP107 |
|---|---|---|
| GPU variant | Fiji XT | GP107-860-A1 |
| Architecture | GCN 3.0 | Pascal |
| Fonderie | TSMC | Samsung |
| Taille du processus | 28 nm | 14 nm |
| Transistors | 8,900 million | 3,300 million |
| Taille des matrices | 596 mm² | 132 mm² |
Carte graphique
| Date de publication | Dec 12th, 2016 | Feb 7th, 2017 |
|---|---|---|
| Génération | Radeon Instinct | Quadro |
| Production | End-of-life | End-of-life |
| Bus interface | PCIe 3.0 x16 | PCIe 3.0 x16 |
Vitesse clock
| GPU clock | 1000 MHz | — |
|---|---|---|
| Mémoire clock | 500 MHz 1000 Mbps effective | 1502 MHz 6 Gbps effective |
| Base clock | — | 1354 MHz |
| Boost clock | — | 1392 MHz |
Mémoire
| Taille de la mémoire | 4 GB | 4 GB |
|---|---|---|
| Type de mémoire | HBM | GDDR5 |
| Bus mémoire | 4096 bit | 128 bit |
| Bande passante | 512.0 GB/s | 96.13 GB/s |
Render Config
| Shading units | 4096 | 512 |
|---|---|---|
| TMUs | 256 | 32 |
| ROPs | 64 | 16 |
| Compute units | 64 | — |
| Cache L1 | 16 KB (per CU) | 48 KB (per SM) |
| Cache L2 | 2 MB | 1024 KB |
| Comptage SM | — | 4 |
Performance théorique
| Taux de pixel | 64.00 GPixel/s | 22.27 GPixel/s |
|---|---|---|
| Taux de texture | 256.0 GTexel/s | 44.54 GTexel/s |
| FP16 (half) performance | 8.192 TFLOPS (1:1) | 22.27 GFLOPS (1:64) |
| FP32 (float) performance | 8.192 TFLOPS | 1,425 GFLOPS |
| FP64 (double) performance | 512.0 GFLOPS (1:16) | 44.54 GFLOPS (1:32) |
Board Design
| Taille Slot | Dual-slot | Single-slot |
|---|---|---|
| Longueur | 152 mm 6 inches | 150 mm 5.9 inches |
| Taille | 111 mm 4.4 inches | 69 mm 2.7 inches |
| TDP | 175 W | 47 W |
| Suggestions pour le PSU | 450 W | 200 W |
| Sorties | No outputs | 4x mini-DisplayPort |
| Connecteurs de puissance | 1x 8-pin | None |
| Board number | — | PG212 SKU 502 |
Caractéristiques graphiques
| DirectX | 12 (12_0) | 12 (12_1) |
|---|---|---|
| OpenGL | 4.6 | 4.6 |
| OpenCL | 2.0 | 3.0 |
| Vulkan | 1.2 | 1.2 |
| Shader model | 6.3 | 6.4 |
| CUDA | — | 6.1 |
