GeForce RTX 3060 Ti GDDR6X vs Radeon HD 8670 OEM
We compared two Desktop platform GPUs: 8GB VRAM GeForce RTX 3060 Ti GDDR6X and 2GB VRAM Radeon HD 8670 OEM to see which GPU has better performance in key specifications, benchmark tests, power consumption, etc.
Main Differences
GeForce RTX 3060 Ti GDDR6X 's Advantages
Released 9 years and 9 months late
Boost Clock has increased by 59% (1665MHz vs 1050MHz)
More VRAM (8GB vs 2GB)
Larger VRAM bandwidth (608.3GB/s vs 73.60GB/s)
4480 additional rendering cores
Radeon HD 8670 OEM 's Advantages
Lower TDP (75W vs 225W)
Score
Benchmark
FP32 (float)
GeForce RTX 3060 Ti GDDR6X
+1909%
16200
Radeon HD 8670 OEM
806
Graphics Card
Oct 2022
Release Date
Jan 2013
GeForce 30
Generation
Sea Islands
Desktop
Type
Desktop
PCIe 4.0 x16
Bus Interface
PCIe 3.0 x8
Clock Speeds
1410 MHz
Base Clock
1000 MHz
1665 MHz
Boost Clock
1050 MHz
1188 MHz
Memory Clock
1150 MHz
Memory
8GB
Memory Size
2GB
GDDR6X
Memory Type
GDDR5
256bit
Memory Bus
128bit
608.3GB/s
Bandwidth
73.60GB/s
Render Config
-
Compute Units
6
38
SM Count
-
4864
Shading Units
384
152
TMUs
24
80
ROPs
8
152
Tensor Cores
-
38
RT Cores
-
128 KB (per SM)
L1 Cache
16 KB (per CU)
4 MB
L2 Cache
256 KB
Theoretical Performance
133.2 GPixel/s
Pixel Rate
8.400 GPixel/s
253.1 GTexel/s
Texture Rate
25.20 GTexel/s
16.20 TFLOPS
FP16 (half)
-
16.20 TFLOPS
FP32 (float)
806.4 GFLOPS
253.1 GFLOPS
FP64 (double)
50.40 GFLOPS
Board Design
225W
TDP
75W
550 W
Suggested PSU
250 W
1x HDMI 2.1
3x DisplayPort 1.4a
Outputs
1x DVI
1x HDMI 1.4a
1x VGA
1x 12-pin
Power Connectors
None
Graphics Processor
GA104
GPU Name
Oland
GA104-202-A1
GPU Variant
Oland XT
(215-0837000)
Ampere
Architecture
GCN 1.0
Samsung
Foundry
TSMC
8 nm
Process Size
28 nm
17.4 billion
Transistors
0.95 billion
392 mm²
Die Size
77 mm²
Graphics Features
12 Ultimate (12_2)
DirectX
12 (11_1)
4.6
OpenGL
4.6
3.0
OpenCL
1.2
1.3
Vulkan
1.2
8.6
CUDA
-
6.7
Shader Model
5.1
