AMD Instinct MI300X vs NVIDIA GeForce RTX 4090
Side-by-side comparison of specifications, performance, and efficiency.
AMD Instinct MI300X
Architecture
CDNA 3
Year
2023
VRAM
192 GB
TDP
750 W
NVIDIA GeForce RTX 4090
Architecture
Ada Lovelace
Year
2022
VRAM
24 GB
TDP
450 W
Specifications
AMD Instinct MI300X NVIDIA GeForce RTX 4090
Architecture
CDNA 3 Ada Lovelace
Launch Year
2023 2022
Form Factor
OAM PCIe
VRAM
192 GB24 GB
Memory Bandwidth
5.3 TB/s1.0 TB/s
TDP
750 W450 W
Process Node
5nm FinFET 4nm
| Spec | AMD Instinct MI300X | NVIDIA GeForce RTX 4090 |
|---|---|---|
| Architecture | CDNA 3 | Ada Lovelace |
| Launch Year | 2023 | 2022 |
| Form Factor | OAM | PCIe |
| VRAM | 192 GB | 24 GB |
| Memory Bandwidth | 5.3 TB/s | 1.0 TB/s |
| TDP | 750 W | 450 W |
| Process Node | 5nm FinFET | 4nm |
Performance (TFLOPS)
AMD Instinct MI300X NVIDIA GeForce RTX 4090
FP64
81.7 TFlops1.3 TFlops
FP32
163.4 TFlops82.6 TFlops
FP16
1307.4 TFlops165.2 TFlops
BF16
1307.4 TFlops165.2 TFlops
FP8
2614.9 TFlops330.3 TFlops
INT8
2614.9 TFlops330.3 TFlops
| Precision | AMD Instinct MI300X | NVIDIA GeForce RTX 4090 |
|---|---|---|
| FP64 | 81.7 TFlops | 1.3 TFlops |
| FP32 | 163.4 TFlops | 82.6 TFlops |
| FP16 | 1307.4 TFlops | 165.2 TFlops |
| BF16 | 1307.4 TFlops | 165.2 TFlops |
| FP8 | 2614.9 TFlops | 330.3 TFlops |
| INT8 | 2614.9 TFlops | 330.3 TFlops |
FLOPS by Precision
Efficiency — TDP vs FP32 Performance
Analysis
Most Efficient
AMD
Instinct MI300X
0.22 TFLOPS/W
Get Comparison Updates
New GPUs added weekly. Be the first to see how they compare.