NVIDIA RTX A4500 vs NVIDIA Tesla C870
Processeur graphique
| Nom du GPU | GA102 | G80 |
|---|---|---|
| Architecture | Ampere | Tesla |
| Fonderie | Samsung | TSMC |
| Taille du processus | 8 nm | 90 nm |
| Transistors | 28,300 million | 681 million |
| Taille des matrices | 628 mm² | 484 mm² |
Carte graphique
| Date de publication | Nov 23rd, 2021 | May 2nd, 2007 |
|---|---|---|
| Génération | Quadro | Tesla |
| Production | Active | End-of-life |
| Bus interface | PCIe 4.0 x16 | PCIe 1.0 x16 |
| Prix de lancement | — | 1,499 USD |
Vitesse clock
| Base clock | 1050 MHz | — |
|---|---|---|
| Boost clock | 1650 MHz | — |
| Mémoire clock | 2000 MHz 16 Gbps effective | 800 MHz 1600 Mbps effective |
| GPU clock | — | 600 MHz |
| Shader clock | — | 1350 MHz |
Mémoire
| Taille de la mémoire | 20 GB | 1536 MB |
|---|---|---|
| Type de mémoire | GDDR6 | GDDR3 |
| Bus mémoire | 320 bit | 384 bit |
| Bande passante | 640.0 GB/s | 76.80 GB/s |
Render Config
| Shading units | 7168 | 128 |
|---|---|---|
| TMUs | 224 | 32 |
| ROPs | 96 | 24 |
| Comptage SM | 56 | 16 |
| Tensor cores | 224 | — |
| RT cores | 56 | — |
| Cache L1 | 128 KB (per SM) | — |
| Cache L2 | 6 MB | 96 KB |
Performance théorique
| Taux de pixel | 158.4 GPixel/s | 14.40 GPixel/s |
|---|---|---|
| Taux de texture | 369.6 GTexel/s | 38.40 GTexel/s |
| FP16 (half) performance | 23.65 TFLOPS (1:1) | — |
| FP32 (float) performance | 23.65 TFLOPS | 345.6 GFLOPS |
| FP64 (double) performance | 739.2 GFLOPS (1:32) | — |
Board Design
| Taille Slot | Dual-slot | Dual-slot |
|---|---|---|
| Longueur | 267 mm 10.5 inches | — |
| Taille | 112 mm 4.4 inches | — |
| TDP | 200 W | 171 W |
| Suggestions pour le PSU | 550 W | 450 W |
| Sorties | 4x DisplayPort | No outputs |
| Connecteurs de puissance | 1x 8-pin | 2x 6-pin |
| Board number | — | P357 |
Caractéristiques graphiques
| DirectX | 12 Ultimate (12_2) | 11.1 (10_0) |
|---|---|---|
| OpenGL | 4.6 | 3.3 |
| OpenCL | 3.0 | 1.1 (1.0) |
| Vulkan | 1.2 | — |
| CUDA | 8.6 | 1.0 |
| Shader model | 6.6 | 4.0 |
