NVIDIA GeForce GT 140 OEM vs NVIDIA GRID A100B
Processeur graphique
| Nom du GPU | G94B | GA100 |
|---|---|---|
| Architecture | Tesla | Ampere |
| Fonderie | TSMC | TSMC |
| Taille du processus | 55 nm | 7 nm |
| Transistors | 505 million | 54,200 million |
| Taille des matrices | 196 mm² | 826 mm² |
Carte graphique
| Date de publication | Mar 10th, 2009 | May 14th, 2020 |
|---|---|---|
| Génération | GeForce 100 | GRID |
| Production | End-of-life | Active |
| Bus interface | PCIe 2.0 x16 | PCIe 4.0 x16 |
| Prédécesseur | GeForce 9 | — |
| Successeur | GeForce 200 | — |
Vitesse clock
| GPU clock | 650 MHz | — |
|---|---|---|
| Shader clock | 1625 MHz | — |
| Mémoire clock | 900 MHz 1800 Mbps effective | 1215 MHz 2.4 Gbps effective |
| Base clock | — | 900 MHz |
| Boost clock | — | 1005 MHz |
Mémoire
| Taille de la mémoire | 1024 MB | 48 GB |
|---|---|---|
| Type de mémoire | GDDR3 | HBM2E |
| Bus mémoire | 256 bit | 6144 bit |
| Bande passante | 57.60 GB/s | 1,866 GB/s |
Render Config
| Shading units | 64 | 6912 |
|---|---|---|
| TMUs | 32 | 432 |
| ROPs | 16 | 192 |
| Comptage SM | 8 | 108 |
| Cache L2 | 64 KB | 48 MB |
| Tensor cores | — | 432 |
| Cache L1 | — | 192 KB (per SM) |
Performance théorique
| Taux de pixel | 10.40 GPixel/s | 193.0 GPixel/s |
|---|---|---|
| Taux de texture | 20.80 GTexel/s | 434.2 GTexel/s |
| FP32 (float) performance | 208.0 GFLOPS | 13.89 TFLOPS |
| FP16 (half) performance | — | 55.57 TFLOPS (4:1) |
| FP64 (double) performance | — | 6.947 TFLOPS (1:2) |
Board Design
| Taille Slot | Single-slot | IGP |
|---|---|---|
| Longueur | 229 mm 9 inches | — |
| TDP | 105 W | 400 W |
| Suggestions pour le PSU | 300 W | 800 W |
| Sorties | 2x DVI1x S-Video | No outputs |
| Connecteurs de puissance | 1x 6-pin | None |
| Board number | P545 | — |
Caractéristiques graphiques
| DirectX | 11.1 (10_0) | — |
|---|---|---|
| OpenGL | 3.3 | — |
| OpenCL | 1.1 | 3.0 |
| Vulkan | — | — |
| CUDA | 1.1 | 8.0 |
| Shader model | 4.0 | — |
