Nom du GPU | Vega 10 | TU104 |
---|---|---|
GPU variant | Vega 10 XT GL (215-0894124) | TU104-895-A1 |
Architecture | GCN 5.0 | Turing |
Fonderie | GlobalFoundries | TSMC |
Taille du processus | 14 nm | 12 nm |
Transistors | 12,500 million | 13,600 million |
Taille des matrices | 495 mm² | 545 mm² |
Date de publication | Aug 8th, 2017 | Sep 13th, 2018 |
---|---|---|
Génération | Radeon Pro | Tesla |
Production | Active | Active |
Prix de lancement | 6,999 USD | — |
Bus interface | PCIe 3.0 x16 | PCIe 3.0 x16 |
Base clock | 1440 MHz | 585 MHz |
---|---|---|
Boost clock | 1500 MHz | 1590 MHz |
Mémoire clock | 945 MHz 1890 Mbps effective | 1250 MHz 10 Gbps effective |
Taille de la mémoire | 16 GB | 16 GB |
---|---|---|
Type de mémoire | HBM2 | GDDR6 |
Bus mémoire | 2048 bit | 256 bit |
Bande passante | 483.8 GB/s | 320.0 GB/s |
Shading units | 4096 | 2560 |
---|---|---|
TMUs | 256 | 160 |
ROPs | 64 | 64 |
Compute units | 64 | — |
Cache L1 | 16 KB (per CU) | 64 KB (per SM) |
Cache L2 | 4 MB | 4 MB |
Comptage SM | — | 40 |
Tensor cores | — | 320 |
RT cores | — | 40 |
Taux de pixel | 96.00 GPixel/s | 101.8 GPixel/s |
---|---|---|
Taux de texture | 384.0 GTexel/s | 254.4 GTexel/s |
FP16 (half) performance | 24.58 TFLOPS (2:1) | 65.13 TFLOPS (8:1) |
FP32 (float) performance | 12.29 TFLOPS | 8.141 TFLOPS |
FP64 (double) performance | 768.0 GFLOPS (1:16) | 254.4 GFLOPS (1:32) |
Taille Slot | Dual-slot | Single-slot |
---|---|---|
Longueur | 267 mm 10.5 inches | 168 mm 6.6 inches |
Taille | 111 mm 4.4 inches | — |
TDP | 260 W | 70 W |
Suggestions pour le PSU | 600 W | 250 W |
Sorties | 6x mini-DisplayPort | No outputs |
Connecteurs de puissance | 1x 6-pin + 1x 8-pin | None |
Board number | — | PG183 SKU 200 |
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 |