GeForce GTX 1070 and GeForce GTX 680 - video card comparison
Passmark - benchmark score GeForce GTX 1070 vs GeForce GTX 680
GeForce GTX 1070
13916
GeForce GTX 680
5721
GeForce GTX 1070
GeForce GTX 680
Architecture
Pascal (2016−2021)
Kepler (2012−2018)
GPU code name
Pascal GP104
GK104
Market segment
Desktop
Desktop
Release date
May 2016
March 2012
Bus support
PCIe 3.0
PCI Express 3.0
Interface
PCIe 3.0 x16
PCIe 3.0 x16
Recommended system power (PSU)
500 Watt
-
Supplementary power connectors
8-pin
Two 6-pin
SLI options
+
+
Boost clock speed
1683 MHz
1058 MHz
Core clock speed
1506 MHz
1006 MHz
CUDA cores
1920
1536
Floating-point performance
6,463 gflops
3,090.4 gflops
Manufacturing process technology
16 nm
28 nm
Number of transistors
7,200 million
3,540 million
Pipelines / CUDA cores
1920
1536
Power consumption (TDP)
150 Watt
195 Watt
Maximum GPU temperature
94 °C
-
Texture fill rate
202.0
128.8 billion/sec
Memory type
GDDR5
GDDR5
Maximum RAM amount
8 GB
2048 MB
Memory bus width
256 Bit
256-bit GDDR5
Memory clock speed
8 GB/s
6000 MHz
Memory bandwidth
256 GB/s
192.2 GB/s
Shared memory
-
-
Display Connectors
DP 1.42, HDMI 2.0b, Dual Link-DVI
One Dual Link DVI-I, One Dual Link DVI-D, One HDMI, One DisplayPort
Multi monitor support
+
4 displays
HDMI
+
+
HDCP
-
+
Maximum VGA resolution
-
2048x1536
G-SYNC support
+
-
Audio input for HDMI
-
Internal
DirectX
12 (12_1)
12 (11_0)
CUDA
+
+
OpenCL
1.2
1.2
OpenGL
4.5
4.2
Shader Model
6.4
5.1
Vulkan
+
1.1.126
Ansel
+
-
GPU Boost
3.0
-
VR Ready
+
-