NVIDIA A40 PCIe vs NVIDIA L20
We compared two Desktop platform GPUs: 48GB VRAM A40 PCIe and 48GB VRAM L20 to see which GPU has better performance in key specifications, benchmark tests, power consumption, etc.
Main Differences
NVIDIA L20 's Advantages
Released 3 years and 1 months late
Boost Clock has increased by 45% (2520MHz vs 1740MHz)
Larger VRAM bandwidth (864.0GB/s vs 695.8GB/s)
1024 additional rendering cores
Lower TDP (275W vs 300W)
Score
Benchmark
FP32 (float)
A40 PCIe
37.42 TFLOPS
L20
+58%
59.35 TFLOPS
Graphics Card
Oct 2020
Release Date
Nov 2023
Tesla
Generation
Tesla Ada
Desktop
Type
Desktop
PCIe 4.0 x16
Bus Interface
PCIe 4.0 x16
Clock Speeds
1305 MHz
Base Clock
1440 MHz
1740 MHz
Boost Clock
2520 MHz
1812 MHz
Memory Clock
2250 MHz
Memory
48GB
Memory Size
48GB
GDDR6
Memory Type
GDDR6
384bit
Memory Bus
384bit
695.8GB/s
Bandwidth
864.0GB/s
Render Config
-
-
-
84
SM Count
92
10752
Shading Units
11776
336
TMUs
368
112
ROPs
128
336
Tensor Cores
368
84
RT Cores
92
128 KB (per SM)
L1 Cache
128 KB (per SM)
6 MB
L2 Cache
96 MB
-
-
-
Theoretical Performance
194.9 GPixel/s
Pixel Rate
322.6 GPixel/s
584.6 GTexel/s
Texture Rate
927.4 GTexel/s
37.42 TFLOPS
FP16 (half)
59.35 TFLOPS
37.42 TFLOPS
FP32 (float)
59.35 TFLOPS
584.6 GFLOPS
FP64 (double)
927.4 GFLOPS
Board Design
300W
TDP
275W
700 W
Suggested PSU
600 W
3x DisplayPort 1.4a
Outputs
4x DisplayPort 1.4a
8-pin EPS
Power Connectors
1x 16-pin
Graphics Processor
GA102
GPU Name
AD102
-
GPU Variant
AD102
Ampere
Architecture
Ada Lovelace
Samsung
Foundry
TSMC
8 nm
Process Size
5 nm
28.3 billion
Transistors
76.3 billion
628 mm²
Die Size
609 mm²
Graphics Features
12 Ultimate (12_2)
DirectX
12 Ultimate (12_2)
4.6
OpenGL
4.6
3.0
OpenCL
3.0
1.3
Vulkan
1.3
8.6
CUDA
8.9
6.6
Shader Model
6.7
