Nom du GPU | GA100 | NV28 |
---|---|---|
Architecture | Ampere | Kelvin |
Fonderie | TSMC | TSMC |
Taille du processus | 7 nm | 150 nm |
Transistors | 54,200 million | 36 million |
Taille des matrices | 826 mm² | 101 mm² |
GPU variant | — | 4200 Go |
Date de publication | Jun 22nd, 2020 | — |
---|---|---|
Génération | Tesla | — |
Production | Active | — |
Bus interface | PCIe 4.0 x16 | — |
GPU clock | 1410 MHz | 200 MHz |
---|---|---|
Mémoire clock | 1215 MHz 2.4 Gbps effective | 200 MHz 400 Mbps effective |
Base clock | 765 MHz | — |
Boost clock | 1410 MHz | — |
Taille de la mémoire | 40 GB | 64 MB |
---|---|---|
Type de mémoire | HBM2E | DDR |
Bus mémoire | 5120 bit | 64 bit |
Bande passante | 1,555 GB/s | 3.200 GB/s |
Shading units | 6912 | — |
---|---|---|
TMUs | 432 | 8 |
ROPs | 160 | 4 |
Comptage SM | 108 | — |
Tensor cores | 432 | — |
Cache L1 | 192 KB (per SM) | — |
Cache L2 | 40 MB | — |
Pixel shaders | — | 4 |
Vertex shaders | — | 2 |
Taux de pixel | 225.6 GPixel/s | 800.0 MPixel/s |
---|---|---|
Taux de texture | 609.1 GTexel/s | 1.600 GTexel/s |
FP16 (half) performance | 77.97 TFLOPS (4:1) | — |
FP32 (float) performance | 19.49 TFLOPS | — |
FP64 (double) performance | 9.746 TFLOPS (1:2) | — |
Vertex rate | — | 100.0 MVertices/s |
Taille Slot | Dual-slot | — |
---|---|---|
Longueur | 267 mm 10.5 inches | — |
TDP | 250 W | unknown |
Suggestions pour le PSU | 600 W | — |
Sorties | No outputs | No outputs |
Connecteurs de puissance | 8-pin EPS | — |
Board number | P1001 SKU 200 | — |
DirectX | — | 8.1 |
---|---|---|
OpenGL | — | 1.5 |
OpenCL | 3.0 | — |
Vulkan | — | — |
CUDA | 8.0 | — |
Shader model | — | — |
Pixel shader | — | 1.3 |
Vertex shader | — | 1.1 |
Date de publication | — | Nov 14th, 2002 |
---|---|---|
Génération | — | GeForce4 Go |
Production | — | End-of-life |
Bus interface | — | AGP 8x |
Critiques | — | 2 in our database |
Prédécesseur | — | GeForce2 Go |
Successeur | — | GeForce FX Go 5 |