Radeon Instinct MI250 vs L40S
We compared two Professional market GPUs: 128GB VRAM Radeon Instinct MI250 and 48GB VRAM L40S to see which GPU has better performance in key specifications, benchmark tests, power consumption, etc.
Main Differences
Radeon Instinct MI250 Advantages
More VRAM (128GB vs 48GB)
Larger VRAM bandwidth (3277GB/s vs 864.0GB/s)
L40S Advantages
Released 11 months late
Boost Clock has increased by 48% (2520MHz vs 1700MHz)
4864 additional rendering cores
Lower TDP (300W vs 500W)
Score
Benchmark
FP32 (float)
Radeon Instinct MI250
45260
L40S
+102%
91610
Graphics Card
Nov 2021
Release Date
Oct 2022
Radeon Instinct
Generation
Tesla Ada
Professional
Type
Professional
PCIe 4.0 x16
Bus Interface
PCIe 4.0 x16
Clock Speeds
1000 MHz
Base Clock
1110 MHz
1700 MHz
Boost Clock
2520 MHz
1600 MHz
Memory Clock
2250 MHz
Memory
128GB
Memory Size
48GB
HBM2e
Memory Type
GDDR6
8192bit
Memory Bus
384bit
3277GB/s
Bandwidth
864.0GB/s
Render Config
208
Compute Units
-
-
SM Count
142
13312
Shading Units
18176
832
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
1414 GTexel/s
Texture Rate
1431 GTexel/s
362.1 TFLOPS
FP16 (half)
91.61 TFLOPS
45.26 TFLOPS
FP32 (float)
91.61 TFLOPS
45.26 TFLOPS
FP64 (double)
1431 GFLOPS
Board Design
500W
TDP
300W
900 W
Suggested PSU
700 W
No outputs
Outputs
1x HDMI 2.1
3x DisplayPort 1.4a
3x DisplayPort 1.4a
2x 8-pin
Power Connectors
1x 16-pin
Graphics Processor
Aldebaran
GPU Name
AD102
Aldebaran
GPU Variant
AD102-???-A1
CDNA 2.0
Architecture
Ada Lovelace
TSMC
Foundry
TSMC
6 nm
Process Size
5 nm
58.2 billion
Transistors
76.3 billion
Unknown
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
