Nom du GPU | Picasso | GA104 |
---|---|---|
Architecture | GCN 5.0 | Ampere |
Fonderie | GlobalFoundries | Samsung |
Taille du processus | 14 nm | 8 nm |
Transistors | 4,940 million | 17,400 million |
Taille des matrices | 210 mm² | 392 mm² |
Date de publication | Jan 6th, 2019 | — |
---|---|---|
Génération | Picasso (Vega Mobile) | — |
Production | Active | — |
Bus interface | IGP | — |
Prédécesseur | Raven Ridge | — |
Successeur | Renoir | — |
Base clock | 300 MHz | 1260 MHz |
---|---|---|
Boost clock | 1200 MHz | 1560 MHz |
Mémoire clock | System Shared | 1375 MHz 11 Gbps effective |
Taille de la mémoire | System Shared | 6 GB |
---|---|---|
Type de mémoire | System Shared | GDDR6 |
Bus mémoire | System Shared | 192 bit |
Bande passante | System Dependent | 264.0 GB/s |
Shading units | 384 | 4096 |
---|---|---|
TMUs | 24 | 128 |
ROPs | 8 | 64 |
Compute units | 6 | — |
Comptage SM | — | 32 |
Tensor cores | — | 128 |
RT cores | — | 32 |
Cache L1 | — | 128 KB (per SM) |
Cache L2 | — | 4 MB |
Taux de pixel | 9.600 GPixel/s | 99.84 GPixel/s |
---|---|---|
Taux de texture | 28.80 GTexel/s | 199.7 GTexel/s |
FP16 (half) performance | 1.843 TFLOPS (2:1) | 12.78 TFLOPS (1:1) |
FP32 (float) performance | 921.6 GFLOPS | 12.78 TFLOPS |
FP64 (double) performance | 57.60 GFLOPS (1:16) | 199.7 GFLOPS (1:64) |
Taille Slot | IGP | — |
---|---|---|
TDP | 15 W | 130 W |
Sorties | No outputs | No outputs |
Connecteurs de puissance | — | None |
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 | — | 8.6 |
Date de publication | — | Unknown |
---|---|---|
Génération | — | Quadro Mobile (Ax000) |
Production | — | Active |
Bus interface | — | PCIe 4.0 x16 |