Nom du GPU | GA100 | GA104 |
---|---|---|
Architecture | Ampere | Ampere |
Fonderie | TSMC | Samsung |
Taille du processus | 7 nm | 8 nm |
Transistors | 54,200 million | 17,400 million |
Taille des matrices | 826 mm² | 392 mm² |
GPU variant | — | GA104-875-A1 |
Date de publication | Jun 22nd, 2020 | Apr 12th, 2021 |
---|---|---|
Génération | Tesla | Quadro |
Production | Active | Active |
Bus interface | PCIe 4.0 x16 | PCIe 4.0 x16 |
GPU clock | 1410 MHz | — |
---|---|---|
Mémoire clock | 1215 MHz 2.4 Gbps effective | 1750 MHz 14 Gbps effective |
Base clock | 765 MHz | 735 MHz |
Boost clock | 1410 MHz | 1560 MHz |
Taille de la mémoire | 40 GB | 16 GB |
---|---|---|
Type de mémoire | HBM2E | GDDR6 |
Bus mémoire | 5120 bit | 256 bit |
Bande passante | 1,555 GB/s | 448.0 GB/s |
Shading units | 6912 | 6144 |
---|---|---|
TMUs | 432 | 192 |
ROPs | 160 | 96 |
Comptage SM | 108 | 48 |
Tensor cores | 432 | 192 |
Cache L1 | 192 KB (per SM) | 128 KB (per SM) |
Cache L2 | 40 MB | 4 MB |
RT cores | — | 48 |
Taux de pixel | 225.6 GPixel/s | 149.8 GPixel/s |
---|---|---|
Taux de texture | 609.1 GTexel/s | 299.5 GTexel/s |
FP16 (half) performance | 77.97 TFLOPS (4:1) | 19.17 TFLOPS (1:1) |
FP32 (float) performance | 19.49 TFLOPS | 19.17 TFLOPS |
FP64 (double) performance | 9.746 TFLOPS (1:2) | 599.0 GFLOPS (1:32) |
Taille Slot | Dual-slot | Single-slot |
---|---|---|
Longueur | 267 mm 10.5 inches | 241 mm 9.5 inches |
TDP | 250 W | 140 W |
Suggestions pour le PSU | 600 W | 300 W |
Sorties | No outputs | 4x DisplayPort |
Connecteurs de puissance | 8-pin EPS | 1x 6-pin |
Board number | P1001 SKU 200 | — |
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 |