GeForce 8800 GTX and Radeon HD 3830 - video card comparison
GeForce 8800 GTX
Radeon HD 3830
Architecture
Tesla (2006−2010)
TeraScale (2005−2013)
GPU code name
G80
RV670
Market segment
Desktop
Desktop
Release date
November 2006
April 2008
Interface
PCIe 1.0 x16
PCIe 2.0 x16
Supplementary power connectors
2x 6-pin
1x 6-pin
SLI options
+
-
Core clock speed
576 MHz
668 MHz
CUDA cores
575
-
Floating-point performance
345.6 gflops
427.5 gflops
Manufacturing process technology
90 nm
55 nm
Number of transistors
681 million
666 million
Pipelines / CUDA cores
128
320
Power consumption (TDP)
155 Watt
75 Watt
Texture fill rate
36.8 billion/sec
10.69
Memory type
GDDR3
GDDR3
Maximum RAM amount
768 MB
256 MB
Memory bus width
384 Bit
128 Bit
Memory clock speed
900 MHz
1656 MHz
Memory bandwidth
86.4 GB/s
26.5 GB/s
Display Connectors
2x DVI, 1x S-Video
No outputs
DirectX
11.1 (10_0)
10.1 (10_1)
CUDA
+
-
OpenCL
1.1
N/A
OpenGL
3.3
3.3
Shader Model
4.0
4.1
Vulkan
N/A
N/A