NVIDIA GM200 Graphics Processing Unit (GPU)

Overview
Manufacturer
NVIDIA
Series
Quadro Maxwell
Release Date
March 19th, 2015
Board Model
NVIDIA PG600
Graphics Processing Unit
GPU Model
GM200
Architecture
Maxwell
Fabrication Process
28 nm
Die Size
601 mm2
Transistors Count
8B
Transistors Density
13.3M TRAN/mm2
CUDA Cores
3072
SMMs
24
GPCs
6
TMUs
192
ROPs
96
Clocks
Base Clock
988 MHz
Boost Clock
1140 MHz
Memory Clock
1653 MHz
Effective Memory Clock
6612 Mbps
Memory Configuration
Memory Size
12288 MB
Memory Type
GDDR5
Memory Bus Width
384-bit
Memory Bandwidth
317.4 GB/s
Physical
Interface
PCI-Express 3.0 x16
Power Connectors
1 × 8-pin
Thermal Design Power
250 W
Recommended PSU
600 W
SLI Support
4-way
Display Outputs
DVI-I
1 ×
DisplayPort
3 ×
HDMI 2.0
1 ×
API Support
DirectX
12.0
Mantle
-
Vulkan
1.0
OpenGL
4.5
OpenCL
1.1
Shader Model
5.0

Performance
Pixel Fillrate
94.8 GPixels/s
Texture Fillrate
189.7 GTexel/s
SPFP Performance
7 TFLOPS
Performance per W
28 GFLOPS/W
Performance per mm2
11.7 GFLOPS/mm2


 ModelGPUClocksMemory
NVIDIA GeForce GTX TITAN X GM200-400 1000 / 1750 MHz 12 GB GDDR5
NVIDIA Quadro M6000 GM200 988 / 1653 MHz 12 GB GDDR5
NVIDIA Quadro VCA GM200 988 / 1653 MHz 12 GB GDDR5
NVIDIA Tesla M40 GM200-895 948 / 1500 MHz 12 GB GDDR5
NVIDIA GeForce GTX 980 Ti GM200-310 1000 / 1750 MHz 6 GB GDDR5
NVIDIA Quadro M6000 (V)
3072 Cores
GM200 988 / 1653 MHz 24 GB GDDR5