Nom du GPU | GP104 | GA106 |
---|---|---|
GPU variant | GP104-140-KA-A1 | GA106-140-A1 |
Architecture | Pascal | Ampere |
Fonderie | TSMC | Samsung |
Taille du processus | 16 nm | 8 nm |
Transistors | 7,200 million | 13,250 million |
Taille des matrices | 314 mm² | 276 mm² |
Date de publication | Dec 25th, 2016 | 2022 |
---|---|---|
Génération | GeForce 10 | GeForce 30 |
Prédécesseur | GeForce 900 | GeForce 20 |
Successeur | GeForce 20 | — |
Production | End-of-life | Unreleased |
Prix de lancement | 199 USD | — |
Bus interface | PCIe 3.0 x16 | PCIe 4.0 x16 |
Critiques | 117 in our database | — |
gpu.details.availability | — | 2022 |
Base clock | 1506 MHz | 1545 MHz |
---|---|---|
Boost clock | 1708 MHz | 1740 MHz |
Mémoire clock | 2002 MHz 8 Gbps effective | 1750 MHz 14 Gbps effective |
Taille de la mémoire | 3 GB | 4 GB |
---|---|---|
Type de mémoire | GDDR5 | GDDR6 |
Bus mémoire | 192 bit | 128 bit |
Bande passante | 192.2 GB/s | 224.0 GB/s |
Shading units | 1152 | 2304 |
---|---|---|
TMUs | 72 | 72 |
ROPs | 48 | 40 |
Comptage SM | 9 | 18 |
Cache L1 | 48 KB (per SM) | 128 KB (per SM) |
Cache L2 | 1536 KB | 2 MB |
Tensor cores | — | 72 |
RT cores | — | 18 |
Taux de pixel | 81.98 GPixel/s | 69.60 GPixel/s |
---|---|---|
Taux de texture | 123.0 GTexel/s | 125.3 GTexel/s |
FP16 (half) performance | 61.49 GFLOPS (1:64) | 8.018 TFLOPS (1:1) |
FP32 (float) performance | 3.935 TFLOPS | 8.018 TFLOPS |
FP64 (double) performance | 123.0 GFLOPS (1:32) | 125.3 GFLOPS (1:64) |
Taille Slot | Dual-slot | Dual-slot |
---|---|---|
Longueur | 250 mm 9.8 inches | 242 mm 9.5 inches |
TDP | 120 W | 90 W |
Suggestions pour le PSU | 300 W | 250 W |
Sorties | 1x DVI1x HDMI3x DisplayPort | 1x HDMI3x DisplayPort |
Connecteurs de puissance | 1x 6-pin | 1x 6-pin |
Taille | — | 112 mm 4.4 inches |
DirectX | 12 (12_1) | 12 Ultimate (12_2) |
---|---|---|
OpenGL | 4.6 | 4.6 |
OpenCL | 3.0 | 3.0 |
Vulkan | 1.2 | 1.2 |
CUDA | 6.1 | 8.6 |
Shader model | 6.4 | 6.6 |