Nom du GPU | Vega 20 | TU102 |
---|---|---|
GPU variant | Vega 20 XT | TU102-875-A1 |
Architecture | GCN 5.1 | Turing |
Fonderie | TSMC | TSMC |
Taille du processus | 7 nm | 12 nm |
Transistors | 13,230 million | 18,600 million |
Taille des matrices | 331 mm² | 754 mm² |
Date de publication | Jun 3rd, 2019 | Aug 13th, 2018 |
---|---|---|
Génération | Radeon Pro Mac | Quadro |
Production | Active | Active |
Bus interface | PCIe 3.0 x16 | PCIe 3.0 x16 |
Prix de lancement | — | 6,299 USD |
Base clock | 1574 MHz | 1305 MHz |
---|---|---|
Boost clock | 1720 MHz | 1560 MHz |
Mémoire clock | 806 MHz 1612 Mbps effective | 1750 MHz 14 Gbps effective |
Taille de la mémoire | 32 GB | 24 GB |
---|---|---|
Type de mémoire | HBM2 | GDDR6 |
Bus mémoire | 4096 bit | 384 bit |
Bande passante | 825.3 GB/s | 672.0 GB/s |
Shading units | 4096 | 4608 |
---|---|---|
TMUs | 256 | 288 |
ROPs | 64 | 96 |
Compute units | 64 | — |
Cache L1 | 16 KB (per CU) | 64 KB (per SM) |
Cache L2 | 4 MB | 6 MB |
Comptage SM | — | 72 |
Tensor cores | — | 576 |
RT cores | — | 72 |
Taux de pixel | 110.1 GPixel/s | 149.8 GPixel/s |
---|---|---|
Taux de texture | 440.3 GTexel/s | 449.3 GTexel/s |
FP16 (half) performance | 28.18 TFLOPS (2:1) | 28.75 TFLOPS (2:1) |
FP32 (float) performance | 14.09 TFLOPS | 14.38 TFLOPS |
FP64 (double) performance | 880.6 GFLOPS (1:16) | 449.3 GFLOPS (1:32) |
Taille Slot | Quad-slot | Dual-slot |
---|---|---|
TDP | 475 W | 260 W |
Suggestions pour le PSU | 850 W | 600 W |
Sorties | 1x HDMI4x Thunderbolt | No outputs |
Connecteurs de puissance | None | 1x 6-pin + 1x 8-pin |
Board number | MW672ZM/A | PG150 SKU 510 |
Longueur | — | 267 mm 10.5 inches |
Taille | — | 111 mm 4.4 inches |
DirectX | 12 (12_1) | 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 |