ATI Radeon HD 4670 vs NVIDIA A40 PCIe
We compared two Desktop platform GPUs: 512MB VRAM Radeon HD 4670 and 48GB VRAM A40 PCIe to see which GPU has better performance in key specifications, benchmark tests, power consumption, etc.
Main Differences
ATI Radeon HD 4670 's Advantages
Lower TDP (59W vs 300W)
NVIDIA A40 PCIe 's Advantages
Released 12 years and 1 months late
Boost Clock1740MHz
More VRAM (48GB vs 512GB)
Larger VRAM bandwidth (695.8GB/s vs 32.00GB/s)
10432 additional rendering cores
Score
Benchmark
FP32 (float)
Radeon HD 4670
0.48 TFLOPS
A40 PCIe
+7695%
37.42 TFLOPS
Graphics Card
Sep 2008
Release Date
Oct 2020
Radeon R700
Generation
Tesla
Desktop
Type
Desktop
PCIe 2.0 x16
Bus Interface
PCIe 4.0 x16
Clock Speeds
-
Base Clock
1305 MHz
-
Boost Clock
1740 MHz
1000 MHz
Memory Clock
1812 MHz
Memory
512MB
Memory Size
48GB
GDDR3
Memory Type
GDDR6
128bit
Memory Bus
384bit
32.00GB/s
Bandwidth
695.8GB/s
Render Config
4
Compute Units
-
-
SM Count
84
320
Shading Units
10752
32
TMUs
336
8
ROPs
112
-
Tensor Cores
336
-
RT Cores
84
16 KB (per CU)
L1 Cache
128 KB (per SM)
128 KB
L2 Cache
6 MB
-
-
-
Theoretical Performance
6.000 GPixel/s
Pixel Rate
194.9 GPixel/s
24.00 GTexel/s
Texture Rate
584.6 GTexel/s
-
FP16 (half)
37.42 TFLOPS
480.0 GFLOPS
FP32 (float)
37.42 TFLOPS
-
FP64 (double)
584.6 GFLOPS
Board Design
59W
TDP
300W
250 W
Suggested PSU
700 W
2x DVI
1x S-Video
Outputs
3x DisplayPort 1.4a
None
Power Connectors
8-pin EPS
Graphics Processor
RV730
GPU Name
GA102
RV730 XT
(215-0719047)
GPU Variant
-
TeraScale
Architecture
Ampere
TSMC
Foundry
Samsung
55 nm
Process Size
8 nm
0.514 billion
Transistors
28.3 billion
146 mm²
Die Size
628 mm²
Graphics Features
10.1 (10_1)
DirectX
12 Ultimate (12_2)
3.3
OpenGL
4.6
1.1
OpenCL
3.0
N/A
Vulkan
1.3
-
CUDA
8.6
4.1
Shader Model
6.6
