AMD Instinct MI300A vs NVIDIA L40S
We compared two Professional market GPUs: 128GB VRAM AMD Instinct MI300A and 48GB VRAM L40S to see which GPU has better performance in key specifications, benchmark tests, power consumption, etc.
Main Differences
AMD Instinct MI300A 's Advantages
Released 1 years and 2 months late
More VRAM (128GB vs 48GB)
Larger VRAM bandwidth (5300GB/s vs 864.0GB/s)
NVIDIA L40S 's Advantages
Boost Clock has increased by 20% (2520MHz vs 2100MHz)
3584 additional rendering cores
Lower TDP (300W vs 760W)
Score
Benchmark
FP32 (float)
AMD Instinct MI300A
+33%
122.6 TFLOPS
L40S
91.61 TFLOPS
Graphics Card
Dec 2023
Release Date
Oct 2022
Instinct
Generation
Tesla Ada
Professional
Type
Professional
PCIe 5.0 x16
Bus Interface
PCIe 4.0 x16
Clock Speeds
1000 MHz
Base Clock
1110 MHz
2100 MHz
Boost Clock
2520 MHz
5200 MHz
Memory Clock
2250 MHz
Memory
128GB
Memory Size
48GB
HBM3
Memory Type
GDDR6
8192bit
Memory Bus
384bit
5300GB/s
Bandwidth
864.0GB/s
Render Config
228
Compute Units
-
-
SM Count
142
14592
Shading Units
18176
880
TMUs
568
0
ROPs
192
-
Tensor Cores
568
-
RT Cores
142
16 KB (per CU)
L1 Cache
128 KB (per SM)
16 MB
L2 Cache
48 MB
-
-
-
Theoretical Performance
0 MPixel/s
Pixel Rate
483.8 GPixel/s
1496 GTexel/s
Texture Rate
1431 GTexel/s
980.6 TFLOPS
FP16 (half)
91.61 TFLOPS
122.6 TFLOPS
FP32 (float)
91.61 TFLOPS
61.3 TFLOPS
FP64 (double)
1431 GFLOPS
Board Design
760W
TDP
300W
1000 W
Suggested PSU
700 W
No outputs
Outputs
1x HDMI 2.1
3x DisplayPort 1.4a
None
Power Connectors
1x 16-pin
Graphics Processor
MI300
GPU Name
AD102
-
GPU Variant
AD102-???-A1
CDNA 3.0
Architecture
Ada Lovelace
TSMC
Foundry
TSMC
5 nm
Process Size
5 nm
146 billion
Transistors
76.3 billion
1017 mm²
Die Size
609 mm²
Graphics Features
N/A
DirectX
12 Ultimate (12_2)
N/A
OpenGL
4.6
3.0
OpenCL
3.0
N/A
Vulkan
1.3
-
CUDA
8.9
N/A
Shader Model
6.7
