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