Nom du GPU | Navi 23 | TU102 |
---|---|---|
Architecture | RDNA 2.0 | Turing |
Fonderie | TSMC | TSMC |
Taille du processus | 7 nm | 12 nm |
Transistors | 11,060 million | 18,600 million |
Taille des matrices | 237 mm² | 754 mm² |
GPU variant | — | TU102-875-A1 |
Date de publication | Jun 8th, 2021 | Aug 13th, 2018 |
---|---|---|
gpu.details.availability | Jul, 2021 | — |
Génération | Radeon Pro | Quadro |
Production | Active | Active |
Prix de lancement | 649 USD | 6,299 USD |
Bus interface | PCIe 4.0 x16 | PCIe 3.0 x16 |
Base clock | 2331 MHz | 1305 MHz |
---|---|---|
Boost clock | 2903 MHz | 1560 MHz |
Mémoire clock | 1750 MHz 14 Gbps effective | 1750 MHz 14 Gbps effective |
Taille de la mémoire | 8 GB | 24 GB |
---|---|---|
Type de mémoire | GDDR6 | GDDR6 |
Bus mémoire | 128 bit | 384 bit |
Bande passante | 224.0 GB/s | 672.0 GB/s |
Shading units | 1792 | 4608 |
---|---|---|
TMUs | 112 | 288 |
ROPs | 64 | 96 |
Compute units | 28 | — |
RT cores | 28 | 72 |
gpu.details.l0-cache | 32 KB per WGP | — |
Cache L1 | 128 KB per Array | 64 KB (per SM) |
Cache L2 | 2 MB | 6 MB |
gpu.details.l3-cache | 32 MB | — |
Comptage SM | — | 72 |
Tensor cores | — | 576 |
Taux de pixel | 185.8 GPixel/s | 149.8 GPixel/s |
---|---|---|
Taux de texture | 325.1 GTexel/s | 449.3 GTexel/s |
FP16 (half) performance | 20.81 TFLOPS (2:1) | 28.75 TFLOPS (2:1) |
FP32 (float) performance | 10.40 TFLOPS | 14.38 TFLOPS |
FP64 (double) performance | 650.3 GFLOPS (1:16) | 449.3 GFLOPS (1:32) |
Taille Slot | Single-slot | Dual-slot |
---|---|---|
Longueur | 241 mm 9.5 inches | 267 mm 10.5 inches |
TDP | 100 W | 260 W |
Suggestions pour le PSU | 300 W | 600 W |
Sorties | 4x DisplayPort | No outputs |
Connecteurs de puissance | 1x 6-pin | 1x 6-pin + 1x 8-pin |
Taille | — | 111 mm 4.4 inches |
Board number | — | PG150 SKU 510 |
DirectX | 12.0 Ultimate (12_2) | 12 Ultimate (12_2) |
---|---|---|
OpenGL | 4.6 | 4.6 |
OpenCL | 2.1 | 3.0 |
Vulkan | 1.2 | 1.2 |
Shader model | 6.5 | 6.6 |
CUDA | — | 7.5 |