GPU Name | Fiji | TU106 |
---|---|---|
GPU Variante | Fiji XT C8 (215-0862040) | TU106-200A-KA-A1 |
Architektur | GCN 3.0 | Turing |
Hersteller | TSMC | TSMC |
Prozessgröße | 28 nm | 12 nm |
Transistoren | 8,900 million | 10,800 million |
Chipgröße | 596 mm² | 445 mm² |
Veröffentlichungsdatum | Jun 24th, 2015 | Jan 7th, 2019 |
---|---|---|
Generation | Pirate Islands | GeForce 20 |
Produktion | End-of-life | Active |
Originalpreis | 649 USD | 349 USD |
Businterface | PCIe 3.0 x16 | PCIe 3.0 x16 |
Vorgänger | Volcanic Islands | GeForce 10 |
Nachfolger | Arctic Islands | GeForce 30 |
Reviews | — | 85 in our database |
GPU-Taktung | 1050 MHz | — |
---|---|---|
Speicher-Taktung | 500 MHz 1000 Mbps effective | 1750 MHz 14 Gbps effective |
Basis-Taktfrequenz | — | 1365 MHz |
Boost-Takt | — | 1680 MHz |
Speichergröße | 4 GB | 6 GB |
---|---|---|
Speichertyp | HBM | GDDR6 |
Speicherbus | 4096 bit | 192 bit |
Bandbreite | 512.0 GB/s | 336.0 GB/s |
Shading-Einheiten | 4096 | 1920 |
---|---|---|
TMUs | 256 | 120 |
ROPs | 64 | 48 |
Compute Units | 64 | — |
Cache L1 | 16 KB (per CU) | 64 KB (per SM) |
Cache L2 | 2 MB | 3 MB |
Anzahl SM | — | 30 |
Tensor Cores | — | 240 |
RT-Kerne | — | 30 |
Pixelrate | 67.20 GPixel/s | 80.64 GPixel/s |
---|---|---|
Texturrate | 268.8 GTexel/s | 201.6 GTexel/s |
FP16 (half) Leistung | 8.602 TFLOPS (1:1) | 12.90 TFLOPS (2:1) |
FP32 (float) Leistung | 8.602 TFLOPS | 6.451 TFLOPS |
FP64 (double) Leistung | 537.6 GFLOPS (1:16) | 201.6 GFLOPS (1:32) |
Steckplatzbreite | Dual-slot | Dual-slot |
---|---|---|
Länge | 195 mm 7.7 inches | 229 mm 9 inches |
Breite | 115 mm 4.5 inches | 113 mm 4.4 inches |
Höhe | 39 mm 1.5 inches | 35 mm 1.4 inches |
TDP | 275 W | 160 W |
Vorgeschlagenes Netzteil | 600 W | 450 W |
Ausgabe | 1x HDMI3x DisplayPort | 1x DVI1x HDMI2x DisplayPort1x USB Type-C |
Stromverbindung | 2x 8-pin | 1x 8-pin |
Board-Nummer | C880-37 | PG160 SKU 42 |
DirectX | 12 (12_0) | 12 Ultimate (12_2) |
---|---|---|
OpenGL | 4.6 | 4.6 |
OpenCL | 2.0 | 3.0 |
Vulkan | 1.2 | 1.2 |
Shader-Modell | 6.3 | 6.6 |
CUDA | — | 7.5 |