Nom du GPU | Cape Verde | GA106 |
---|---|---|
GPU variant | Cape Verde PRO MOCHA | GA106-300-A1 |
Architecture | GCN 1.0 | Ampere |
Fonderie | TSMC | Samsung |
Taille du processus | 28 nm | 8 nm |
Transistors | 1,500 million | 13,250 million |
Taille des matrices | 123 mm² | 276 mm² |
Date de publication | Jul 6th, 2016 | Jan 12th, 2021 |
---|---|---|
Génération | Pirate Islands | GeForce 30 |
Production | End-of-life | Active |
Bus interface | PCIe 3.0 x16 | PCIe 4.0 x16 |
Critiques | 61 in our database | 60 in our database |
Prédécesseur | Volcanic Islands | GeForce 20 |
Successeur | Arctic Islands | — |
gpu.details.availability | — | Feb 25th, 2021 |
Prix de lancement | — | 329 USD |
GPU clock | 800 MHz | — |
---|---|---|
Mémoire clock | 1125 MHz 4.5 Gbps effective | 1875 MHz 15 Gbps effective |
Base clock | — | 1320 MHz |
Boost clock | — | 1777 MHz |
Taille de la mémoire | 2 GB | 12 GB |
---|---|---|
Type de mémoire | GDDR5 | GDDR6 |
Bus mémoire | 128 bit | 192 bit |
Bande passante | 72.00 GB/s | 360.0 GB/s |
Shading units | 512 | 3584 |
---|---|---|
TMUs | 32 | 112 |
ROPs | 16 | 48 |
Compute units | 8 | — |
Cache L1 | 16 KB (per CU) | 128 KB (per SM) |
Cache L2 | 256 KB | 3 MB |
Comptage SM | — | 28 |
Tensor cores | — | 112 |
RT cores | — | 28 |
Taux de pixel | 12.80 GPixel/s | 85.30 GPixel/s |
---|---|---|
Taux de texture | 25.60 GTexel/s | 199.0 GTexel/s |
FP32 (float) performance | 819.2 GFLOPS | 12.74 TFLOPS |
FP64 (double) performance | 51.20 GFLOPS (1:16) | 199.0 GFLOPS (1:64) |
FP16 (half) performance | — | 12.74 TFLOPS (1:1) |
Taille Slot | Single-slot | Dual-slot |
---|---|---|
Longueur | 168 mm 6.6 inches | 242 mm 9.5 inches |
TDP | 55 W | 170 W |
Suggestions pour le PSU | 250 W | 450 W |
Sorties | 1x DVI1x HDMI1x DisplayPort | 1x HDMI3x DisplayPort |
Connecteurs de puissance | None | 1x 12-pin |
Taille | — | 112 mm 4.4 inches |
Board number | — | PG190 SKU 60 |
DirectX | 12 (11_1) | 12 Ultimate (12_2) |
---|---|---|
OpenGL | 4.6 | 4.6 |
OpenCL | 1.2 | 3.0 |
Vulkan | 1.2 | 1.2 |
Shader model | 5.1 | 6.6 |
CUDA | — | 8.6 |