Nom du GPU | GA100 | GA102 |
---|---|---|
Architecture | Ampere | Ampere |
Fonderie | TSMC | Samsung |
Taille du processus | 7 nm | 8 nm |
Transistors | 54,200 million | 28,300 million |
Taille des matrices | 826 mm² | 628 mm² |
Date de publication | May 14th, 2020 | Oct 5th, 2020 |
---|---|---|
Génération | Tesla | Quadro |
Production | Active | Active |
Bus interface | PCIe 4.0 x16 | PCIe 4.0 x16 |
gpu.details.availability | — | Dec 15th, 2020 |
Prix de lancement | — | 4,649 USD |
GPU clock | 1410 MHz | — |
---|---|---|
Mémoire clock | 1215 MHz 2.4 Gbps effective | 2000 MHz 16 Gbps effective |
Base clock | 1095 MHz | 1410 MHz |
Boost clock | 1410 MHz | 1800 MHz |
Taille de la mémoire | 40 GB | 48 GB |
---|---|---|
Type de mémoire | HBM2E | GDDR6 |
Bus mémoire | 5120 bit | 384 bit |
Bande passante | 1,555 GB/s | 768.0 GB/s |
Shading units | 6912 | 10752 |
---|---|---|
TMUs | 432 | 336 |
ROPs | 160 | 112 |
Comptage SM | 108 | 84 |
Tensor cores | 432 | 336 |
Cache L1 | 192 KB (per SM) | 128 KB (per SM) |
Cache L2 | 40 MB | 6 MB |
RT cores | — | 84 |
Taux de pixel | 225.6 GPixel/s | 201.6 GPixel/s |
---|---|---|
Taux de texture | 609.1 GTexel/s | 604.8 GTexel/s |
FP16 (half) performance | 77.97 TFLOPS (4:1) | 38.71 TFLOPS (1:1) |
FP32 (float) performance | 19.49 TFLOPS | 38.71 TFLOPS |
FP64 (double) performance | 9.746 TFLOPS (1:2) | 1,210 GFLOPS (1:32) |
Taille Slot | IGP | Dual-slot |
---|---|---|
TDP | 400 W | 300 W |
Suggestions pour le PSU | 800 W | 700 W |
Sorties | No outputs | 4x DisplayPort |
Connecteurs de puissance | None | 8-pin EPS |
Longueur | — | 267 mm 10.5 inches |
Taille | — | 112 mm 4.4 inches |
DirectX | — | 12 Ultimate (12_2) |
---|---|---|
OpenGL | — | 4.6 |
OpenCL | 3.0 | 3.0 |
Vulkan | — | 1.2 |
CUDA | 8.0 | 8.6 |
Shader model | — | 6.6 |