Nom du GPU | Baffin | TU102 |
---|---|---|
GPU variant | Baffin LE | TU102-875-A1 |
Architecture | GCN 4.0 | Turing |
Fonderie | GlobalFoundries | TSMC |
Taille du processus | 14 nm | 12 nm |
Transistors | 3,000 million | 18,600 million |
Taille des matrices | 123 mm² | 754 mm² |
Date de publication | Oct 13th, 2017 | Unknown |
---|---|---|
Génération | Polaris | GRID |
Prédécesseur | Arctic Islands | — |
Successeur | Vega | — |
Production | Active | Active |
Prix de lancement | 79 USD | — |
Bus interface | PCIe 3.0 x8 | PCIe 3.0 x16 |
Base clock | 1019 MHz | 1065 MHz |
---|---|---|
Boost clock | 1071 MHz | 1395 MHz |
Mémoire clock | 1500 MHz 6 Gbps effective | 1750 MHz 14 Gbps effective |
Taille de la mémoire | 2 GB | 2 GB |
---|---|---|
Type de mémoire | GDDR5 | GDDR6 |
Bus mémoire | 128 bit | 384 bit |
Bande passante | 96.00 GB/s | 672.0 GB/s |
Shading units | 512 | 4608 |
---|---|---|
TMUs | 32 | 288 |
ROPs | 16 | 96 |
Compute units | 8 | — |
Cache L1 | 16 KB (per CU) | 64 KB (per SM) |
Cache L2 | 512 KB | 6 MB |
Comptage SM | — | 72 |
Tensor cores | — | 576 |
RT cores | — | 72 |
Taux de pixel | 17.14 GPixel/s | 133.9 GPixel/s |
---|---|---|
Taux de texture | 34.27 GTexel/s | 401.8 GTexel/s |
FP16 (half) performance | 1,097 GFLOPS (1:1) | 25.71 TFLOPS (2:1) |
FP32 (float) performance | 1,097 GFLOPS | 12.86 TFLOPS |
FP64 (double) performance | 68.54 GFLOPS (1:16) | 401.8 GFLOPS (1:32) |
Taille Slot | Dual-slot | Dual-slot |
---|---|---|
Longueur | 145 mm 5.7 inches | 267 mm 10.5 inches |
TDP | 50 W | 260 W |
Suggestions pour le PSU | 250 W | 600 W |
Sorties | 1x DVI1x HDMI1x DisplayPort | No outputs |
Connecteurs de puissance | None | 1x 6-pin + 1x 8-pin |
Board number | D090-01 | PG150 SKU 215 |
DirectX | 12 (12_0) | 12 Ultimate (12_2) |
---|---|---|
OpenGL | 4.6 | 4.6 |
OpenCL | 2.1 | 3.0 |
Vulkan | 1.2 | 1.2 |
Shader model | 6.4 | 6.6 |
CUDA | — | 7.5 |