Architecture
NVIDIA Pascal 16nm
Most Powerful Variant
GP102-875
Cores : TMUs : ROPs
3840 : 240 : 96
Typical Memory Size
24576 MB
Typical Memory Bus
384 -bit
GP102 GPU
Overview
Manufacturer
NVIDIA
Release Date
October 4th, 2016
Graphics Processing Unit
Architecture
Pascal
Fabrication Process
16 nm FinFET
Die Size
471 mm2
Transistors Count
12B
CUDAs
3840
TMUs
240
ROPs
96
Clocks
Base Clock
1506 MHz
Memory Configuration
Memory Size
24576 MB
Memory Type
GDDR5X
Memory Bus Width
384-bit
Memory Bandwidth
432.8 GB/s
Performance
Pixel Fillrate
157.9 GPixels/s
Texture Fillrate
394.8 GTexel/s
SPFP Performance
12.6 TFLOPS
Performance per W
50.5 GFLOPS/W
Performance per mm2
26.8 GFLOPS/mm2
Physical
Interface
PCI-Express 3.0 x16
Typical Thermal Design Power
250 W
API Support
DirectX
12.0
Vulkan
1.0
OpenGL
4.5
OpenCL
3.0
| Model | GPU Model | Unified Cores | Boost Clock | Memory Clock | Memory Config. | |
|---|---|---|---|---|---|---|
 | NVIDIA Quadro P6000 | GP102-875 | ||||
 | NVIDIA Tesla P40 | GP102-895 | ||||
 | NVIDIA TITAN Xp | GP102-450 | ||||
 | NVIDIA TITAN Xp Collector's Edition Star Wars Galactic Empire | GP102-450 | ||||
 | NVIDIA TITAN Xp Collector's Edition Star Wars Jedi Order | GP102-450 | ||||
 | NVIDIA TITAN X (Pascal) | GP102-400 | ||||
 | NVIDIA GeForce GTX 1080 Ti | GP102-350 |
