Nom du GPU | GA102 | GA100 |
---|---|---|
GPU variant | GA102-890-A1 | — |
Architecture | Ampere | Ampere |
Fonderie | Samsung | TSMC |
Taille du processus | 8 nm | 7 nm |
Transistors | 28,300 million | 54,200 million |
Taille des matrices | 628 mm² | 826 mm² |
Date de publication | Apr 12th, 2021 | May 14th, 2020 |
---|---|---|
gpu.details.availability | 2021 | — |
Génération | Tesla | Tesla |
Production | Active | Active |
Bus interface | PCIe 4.0 x16 | PCIe 4.0 x16 |
Base clock | 885 MHz | 1095 MHz |
---|---|---|
Boost clock | 1695 MHz | 1410 MHz |
Mémoire clock | 1563 MHz 12.5 Gbps effective | 1215 MHz 2.4 Gbps effective |
GPU clock | — | 1410 MHz |
Taille de la mémoire | 24 GB | 40 GB |
---|---|---|
Type de mémoire | GDDR6 | HBM2E |
Bus mémoire | 384 bit | 5120 bit |
Bande passante | 600.2 GB/s | 1,555 GB/s |
Shading units | 9216 | 6912 |
---|---|---|
TMUs | 288 | 432 |
ROPs | 96 | 160 |
Comptage SM | 72 | 108 |
Tensor cores | 288 | 432 |
RT cores | 72 | — |
Cache L1 | 128 KB (per SM) | 192 KB (per SM) |
Cache L2 | 6 MB | 40 MB |
Taux de pixel | 162.7 GPixel/s | 225.6 GPixel/s |
---|---|---|
Taux de texture | 488.2 GTexel/s | 609.1 GTexel/s |
FP16 (half) performance | 31.24 TFLOPS (1:1) | 77.97 TFLOPS (4:1) |
FP32 (float) performance | 31.24 TFLOPS | 19.49 TFLOPS |
FP64 (double) performance | 976.3 GFLOPS (1:32) | 9.746 TFLOPS (1:2) |
Taille Slot | Single-slot | IGP |
---|---|---|
Longueur | 267 mm 10.5 inches | — |
Taille | 112 mm 4.4 inches | — |
TDP | 150 W | 400 W |
Suggestions pour le PSU | 450 W | 800 W |
Sorties | No outputs | No outputs |
Connecteurs de puissance | 8-pin EPS | None |
Board number | PG133 SKU 215 | — |
DirectX | 12 Ultimate (12_2) | — |
---|---|---|
OpenGL | 4.6 | — |
OpenCL | 3.0 | 3.0 |
Vulkan | 1.2 | — |
CUDA | 8.6 | 8.0 |
Shader model | 6.6 | — |