AMD Radeon Instinct MI8 vs NVIDIA L4
We compared two Professional market GPUs: 4GB VRAM Radeon Instinct MI8 and 24GB VRAM L4 to see which GPU has better performance in key specifications, benchmark tests, power consumption, etc.
Main Differences
AMD Radeon Instinct MI8 's Advantages
Larger VRAM bandwidth (512.0GB/s vs 300.1GB/s)
NVIDIA L4 's Advantages
Released 6 years and 3 months late
Boost Clock2040MHz
More VRAM (24GB vs 4GB)
3328 additional rendering cores
Lower TDP (72W vs 175W)
Score
Benchmark
FP32 (float)
Radeon Instinct MI8
8.192 TFLOPS
L4
+269%
30.29 TFLOPS
Graphics Card
Dec 2016
Release Date
Mar 2023
Radeon Instinct
Generation
Tesla Ada(Lxx)
Professional
Type
Professional
PCIe 3.0 x16
Bus Interface
PCIe 4.0 x16
Clock Speeds
-
Base Clock
795 MHz
-
Boost Clock
2040 MHz
500 MHz
Memory Clock
1563 MHz
Memory
4GB
Memory Size
24GB
HBM
Memory Type
GDDR6
4096bit
Memory Bus
192bit
512.0GB/s
Bandwidth
300.1GB/s
Render Config
64
Compute Units
-
-
SM Count
60
4096
Shading Units
7424
256
TMUs
240
64
ROPs
80
-
Tensor Cores
240
-
RT Cores
60
16 KB (per CU)
L1 Cache
128 KB (per SM)
2 MB
L2 Cache
48 MB
-
-
-
Theoretical Performance
64.00 GPixel/s
Pixel Rate
163.2 GPixel/s
256.0 GTexel/s
Texture Rate
489.6 GTexel/s
8.192 TFLOPS
FP16 (half)
30.29 TFLOPS
8.192 TFLOPS
FP32 (float)
30.29 TFLOPS
512.0 GFLOPS
FP64 (double)
473.3 GFLOPS
Board Design
175W
TDP
72W
450 W
Suggested PSU
250 W
No outputs
Outputs
No outputs
1x 8-pin
Power Connectors
None
Graphics Processor
Fiji
GPU Name
AD104
Fiji XT CA
(215-0862120)
GPU Variant
AD104-???-A1
GCN 3.0
Architecture
Ada Lovelace
TSMC
Foundry
TSMC
28 nm
Process Size
5 nm
8.9 billion
Transistors
35.8 billion
596 mm²
Die Size
294 mm²
Graphics Features
12 (12_0)
DirectX
12 Ultimate (12_2)
4.6
OpenGL
4.6
2.1
OpenCL
3.0
1.2.170
Vulkan
1.3
-
CUDA
8.9
6.5
Shader Model
6.8
