NVIDIA GRID M40 vs NVIDIA Tesla M40 24 GB
Processeur graphique
| Nom du GPU | GM107 | GM200 |
|---|---|---|
| Architecture | Maxwell | Maxwell 2.0 |
| Fonderie | TSMC | TSMC |
| Taille du processus | 28 nm | 28 nm |
| Transistors | 1,870 million | 8,000 million |
| Taille des matrices | 148 mm² | 601 mm² |
| GPU variant | — | GM200-895-A1 |
Carte graphique
| Date de publication | May 18th, 2016 | Nov 10th, 2015 |
|---|---|---|
| Génération | GRID | Tesla |
| Production | End-of-life | End-of-life |
| Bus interface | PCIe 3.0 x16 | PCIe 3.0 x16 |
Vitesse clock
| GPU clock | 1033 MHz | — |
|---|---|---|
| Mémoire clock | 1300 MHz 5.2 Gbps effective | 1502 MHz 6 Gbps effective |
| Base clock | — | 948 MHz |
| Boost clock | — | 1112 MHz |
Mémoire
| Taille de la mémoire | 8 GB | 24 GB |
|---|---|---|
| Type de mémoire | GDDR5 | GDDR5 |
| Bus mémoire | 128 bit | 384 bit |
| Bande passante | 83.20 GB/s | 288.4 GB/s |
Render Config
| Shading units | 384 | 3072 |
|---|---|---|
| TMUs | 32 | 192 |
| ROPs | 16 | 96 |
| Comptage SMM | 3 | 24 |
| Cache L1 | 64 KB (per SMM) | 48 KB (per SMM) |
| Cache L2 | 2 MB | 3 MB |
Performance théorique
| Taux de pixel | 16.53 GPixel/s | 106.8 GPixel/s |
|---|---|---|
| Taux de texture | 33.06 GTexel/s | 213.5 GTexel/s |
| FP32 (float) performance | 793.3 GFLOPS | 6.832 TFLOPS |
| FP64 (double) performance | 24.79 GFLOPS (1:32) | 213.5 GFLOPS (1:32) |
Board Design
| Taille Slot | Dual-slot | Dual-slot |
|---|---|---|
| TDP | 50 W | 250 W |
| Suggestions pour le PSU | 250 W | 600 W |
| Sorties | No outputs | No outputs |
| Longueur | — | 267 mm 10.5 inches |
| Connecteurs de puissance | — | 1x 6-pin + 1x 8-pin |
| Board number | — | PG600 SKU 202 |
Caractéristiques graphiques
| DirectX | 12 (11_0) | 12 (12_1) |
|---|---|---|
| OpenGL | 4.6 | 4.6 |
| OpenCL | 3.0 | 3.0 |
| Vulkan | 1.1 | 1.1 |
| CUDA | 5.0 | 5.2 |
| Shader model | 5.1 | 6.4 |
