NVIDIA A2 vs NVIDIA Tesla P4
Processeur graphique
| Nom du GPU | GA107 | GP104 |
|---|---|---|
| Architecture | Ampere | Pascal |
| Fonderie | Samsung | TSMC |
| Taille du processus | 8 nm | 16 nm |
| Transistors | unknown | 7,200 million |
| Taille des matrices | unknown | 314 mm² |
| GPU variant | — | GP104-895-A1 |
Carte graphique
| Date de publication | Nov 10th, 2021 | Sep 13th, 2016 |
|---|---|---|
| Génération | Quadro | Tesla |
| Production | Active | End-of-life |
| Bus interface | PCIe 4.0 x8 | PCIe 3.0 x16 |
Vitesse clock
| Base clock | 1440 MHz | 886 MHz |
|---|---|---|
| Boost clock | 1770 MHz | 1114 MHz |
| Mémoire clock | 1563 MHz 12.5 Gbps effective | 1502 MHz 6 Gbps effective |
Mémoire
| Taille de la mémoire | 16 GB | 8 GB |
|---|---|---|
| Type de mémoire | GDDR6 | GDDR5 |
| Bus mémoire | 128 bit | 256 bit |
| Bande passante | 200.1 GB/s | 192.3 GB/s |
Render Config
| Shading units | 1280 | 2560 |
|---|---|---|
| TMUs | 40 | 160 |
| ROPs | 32 | 64 |
| Comptage SM | 10 | 20 |
| Tensor cores | 40 | — |
| RT cores | 10 | — |
| Cache L1 | 128 KB (per SM) | 48 KB (per SM) |
| Cache L2 | 2 MB | 2 MB |
Performance théorique
| Taux de pixel | 56.64 GPixel/s | 71.30 GPixel/s |
|---|---|---|
| Taux de texture | 70.80 GTexel/s | 178.2 GTexel/s |
| FP16 (half) performance | 4.531 TFLOPS (1:1) | 89.12 GFLOPS (1:64) |
| FP32 (float) performance | 4.531 TFLOPS | 5.704 TFLOPS |
| FP64 (double) performance | 70.80 GFLOPS (1:64) | 178.2 GFLOPS (1:32) |
Board Design
| Taille Slot | Single-slot | Single-slot |
|---|---|---|
| TDP | 60 W | 75 W |
| Suggestions pour le PSU | 250 W | 250 W |
| Sorties | No outputs | No outputs |
| Connecteurs de puissance | None | None |
| Board number | PG179 SKU 220 | — |
| Longueur | — | 168 mm 6.6 inches |
Caractéristiques graphiques
| DirectX | 12 Ultimate (12_2) | 12 (12_1) |
|---|---|---|
| OpenGL | 4.6 | 4.6 |
| OpenCL | 3.0 | 3.0 |
| Vulkan | 1.2 | 1.2 |
| CUDA | 8.6 | 6.1 |
| Shader model | 6.6 | 6.4 |
