Nom du GPU | Vega 10 | GV100 |
---|---|---|
GPU variant | Vega 10 XT (215-0894200) | — |
Architecture | GCN 5.0 | Volta |
Fonderie | GlobalFoundries | TSMC |
Taille du processus | 14 nm | 12 nm |
Transistors | 12,500 million | 21,100 million |
Taille des matrices | 495 mm² | 815 mm² |
Date de publication | Jul 10th, 2017 | Nov 26th, 2019 |
---|---|---|
Génération | Radeon Pro | Tesla |
Production | Active | End-of-life |
Prix de lancement | 1,599 USD | — |
Bus interface | PCIe 3.0 x16 | PCIe 3.0 x16 |
Base clock | 1200 MHz | 1245 MHz |
---|---|---|
Boost clock | 1500 MHz | 1597 MHz |
Mémoire clock | 945 MHz 1890 Mbps effective | 1106 MHz 2.2 Gbps effective |
Taille de la mémoire | 16 GB | 32 GB |
---|---|---|
Type de mémoire | HBM2 | HBM2 |
Bus mémoire | 2048 bit | 4096 bit |
Bande passante | 483.8 GB/s | 1,133 GB/s |
Shading units | 4096 | 5120 |
---|---|---|
TMUs | 256 | 320 |
ROPs | 64 | 128 |
Compute units | 64 | — |
Cache L1 | 16 KB (per CU) | 128 KB (per SM) |
Cache L2 | 4 MB | 6 MB |
Comptage SM | — | 80 |
Tensor cores | — | 640 |
Taux de pixel | 96.00 GPixel/s | 204.4 GPixel/s |
---|---|---|
Taux de texture | 384.0 GTexel/s | 511.0 GTexel/s |
FP16 (half) performance | 24.58 TFLOPS (2:1) | 32.71 TFLOPS (2:1) |
FP32 (float) performance | 12.29 TFLOPS | 16.35 TFLOPS |
FP64 (double) performance | 768.0 GFLOPS (1:16) | 8.177 TFLOPS (1:2) |
Taille Slot | Dual-slot | Dual-slot |
---|---|---|
Longueur | 267 mm 10.5 inches | — |
Taille | 111 mm 4.4 inches | — |
TDP | 230 W | 250 W |
Suggestions pour le PSU | 550 W | 600 W |
Sorties | 6x mini-DisplayPort | No outputs |
Connecteurs de puissance | 1x 6-pin + 1x 8-pin | 2x 8-pin |
Board number | D051 | PG500 |
DirectX | 12 (12_1) | 12 (12_1) |
---|---|---|
OpenGL | 4.6 | 4.6 |
OpenCL | 2.1 | 3.0 |
Vulkan | 1.2 | 1.2 |
Shader model | 6.4 | 6.6 |
CUDA | — | 7.0 |