GPU Name | Kaby Lake GT3e | GV10B |
---|---|---|
Architektur | Generation 9.5 | Volta |
Hersteller | Intel | TSMC |
Prozessgröße | 14 nm++ | 12 nm |
Transistoren | unknown | 9,000 million |
Chipgröße | unknown | 350 mm² |
Veröffentlichungsdatum | Jan 3rd, 2017 | Oct 2018 |
---|---|---|
Generation | HD Graphics-M (Kaby Lake) | Tegra |
Produktion | Active | End-of-life |
Businterface | Ring Bus | IGP |
Reviews | 1 in our database | — |
Originalpreis | — | 899 USD |
Basis-Taktfrequenz | 300 MHz | 854 MHz |
---|---|---|
Boost-Takt | 1150 MHz | 1377 MHz |
Speicher-Taktung | System Shared | System Shared |
Speichergröße | System Shared | System Shared |
---|---|---|
Speichertyp | System Shared | System Shared |
Speicherbus | System Shared | System Shared |
Bandbreite | System Dependent | System Dependent |
Shading-Einheiten | 384 | 512 |
---|---|---|
TMUs | 48 | 32 |
ROPs | 6 | 16 |
Execution Units | 48 | — |
Anzahl SM | — | 8 |
Tensor Cores | — | 64 |
Cache L1 | — | 128 KB (per SM) |
Cache L2 | — | 512 KB |
Pixelrate | 6.900 GPixel/s | 22.03 GPixel/s |
---|---|---|
Texturrate | 55.20 GTexel/s | 44.06 GTexel/s |
FP16 (half) Leistung | 1.766 TFLOPS (2:1) | 2.820 TFLOPS (2:1) |
FP32 (float) Leistung | 883.2 GFLOPS | 1,410 GFLOPS |
FP64 (double) Leistung | 220.8 GFLOPS (1:4) | 705.0 GFLOPS (1:2) |
Steckplatzbreite | IGP | IGP |
---|---|---|
TDP | 15 W | 30 W |
Ausgabe | No outputs | No outputs |
Länge | — | 100 mm 3.9 inches |
Breite | — | 87 mm 3.4 inches |
DirectX | 12 (12_1) | 12 (12_1) |
---|---|---|
OpenGL | 4.6 | 4.6 |
OpenCL | 3.0 | 1.2 |
Vulkan | 1.2 | 1.2 |
Shader-Modell | 6.4 | 6.4 |
CUDA | — | 7.2 |