Nom du GPU | G92B | TU106 |
---|---|---|
GPU variant | G92-426-B1 | TU106-400A-A1 |
Architecture | Tesla | Turing |
Fonderie | TSMC | TSMC |
Taille du processus | 55 nm | 12 nm |
Transistors | 754 million | 10,800 million |
Taille des matrices | 260 mm² | 445 mm² |
Date de publication | Mar 4th, 2009 | Oct 17th, 2018 |
---|---|---|
Génération | GeForce 200 | GeForce 20 |
Production | End-of-life | Active |
Prix de lancement | 199 USD | 499 USD |
Bus interface | PCIe 2.0 x16 | PCIe 3.0 x16 |
Critiques | 58 in our database | 73 in our database |
Prédécesseur | GeForce 9 | GeForce 10 |
Successeur | GeForce 400 | GeForce 30 |
GPU clock | 675 MHz | — |
---|---|---|
Shader clock | 1620 MHz | — |
Mémoire clock | 1008 MHz 2 Gbps effective | 1750 MHz 14 Gbps effective |
Base clock | — | 1410 MHz |
Boost clock | — | 1620 MHz |
Taille de la mémoire | 1024 MB | 8 GB |
---|---|---|
Type de mémoire | GDDR3 | GDDR6 |
Bus mémoire | 256 bit | 256 bit |
Bande passante | 64.51 GB/s | 448.0 GB/s |
Shading units | 128 | 2304 |
---|---|---|
TMUs | 64 | 144 |
ROPs | 16 | 64 |
Comptage SM | 16 | 36 |
Cache L2 | 64 KB | 4 MB |
Tensor cores | — | 288 |
RT cores | — | 36 |
Cache L1 | — | 64 KB (per SM) |
Taux de pixel | 10.80 GPixel/s | 103.7 GPixel/s |
---|---|---|
Taux de texture | 43.20 GTexel/s | 233.3 GTexel/s |
FP32 (float) performance | 414.7 GFLOPS | 7.465 TFLOPS |
FP16 (half) performance | — | 14.93 TFLOPS (2:1) |
FP64 (double) performance | — | 233.3 GFLOPS (1:32) |
Taille Slot | Dual-slot | Dual-slot |
---|---|---|
Longueur | 229 mm 9 inches | 229 mm 9 inches |
TDP | 150 W | 175 W |
Suggestions pour le PSU | 450 W | 450 W |
Sorties | 2x DVI | 1x DVI1x HDMI2x DisplayPort1x USB Type-C |
Connecteurs de puissance | 1x 6-pin | 1x 8-pin |
Board number | P365 | PG160 SKU 52 |
Taille | — | 113 mm 4.4 inches |
Taille | — | 35 mm 1.4 inches |
DirectX | 11.1 (10_0) | 12 Ultimate (12_2) |
---|---|---|
OpenGL | 3.3 | 4.6 |
OpenCL | 1.1 | 3.0 |
Vulkan | — | 1.2 |
CUDA | 1.1 | 7.5 |
Shader model | 4.0 | 6.6 |