Compute Comparison
vs
All comparisons →

Ada Lovelace

NVIDIA RTX 4080

wins

3

The RTX 4080 offers 16GB GDDR6X and Ada Lovelace efficiency at a lower price than the RTX 4090.

Budget fine-tuningSmall model inferenceConsumer AI

Ampere

NVIDIA RTX 3090

wins

2

The RTX 3090 remains a popular budget option for AI workloads with 24GB GDDR6X.

Budget fine-tuningSmall model inferenceHobbyist AI

Performance comparison

Visual bars — winner highlighted

NVIDIA RTX 4080

Metric

NVIDIA RTX 3090

FP16 TFLOPS

97 T
71 T

Bandwidth

717 GB/s
936 GB/s

VRAM

166 GB
246 GB

TDP (lower=better)

320 W
350 W

TFLOPS/watt

0.30
0.20

GB/s per watt

2.24
2.67

Full specification table

SpecNVIDIA RTX 4080NVIDIA RTX 3090
ArchitectureAda LovelaceAmpere
VRAM16 GB GDDR6X24 GB GDDR6X
VRAM typeGDDR6XGDDR6X
Memory bandwidth717 GB/s936 GB/s
FP16 TFLOPS97 TFLOPS71 TFLOPS
TDP320 W350 W
TFLOPS/watt0.30 T/W0.20 T/W
GB/s per watt2.24 GB/s/W2.67 GB/s/W
NVLinkNoNo
Release year20222020

Live cloud pricing

On-demand $/hr across providers — updated in real time

Loading prices…

Power efficiency analysis

TFLOPS/watt and GB/s/watt — critical for data center TCO

NVIDIA RTX 4080

FP16 TFLOPS97 TFLOPS
TDP320 W
TFLOPS/watt0.303 T/W
GB/s per watt2.24 GB/s/W

NVIDIA RTX 3090

FP16 TFLOPS71 TFLOPS
TDP350 W
TFLOPS/watt0.203 T/W
GB/s per watt2.67 GB/s/W

TFLOPS/watt measures compute efficiency — how much AI throughput you get per watt of power consumed. For data centers with PUE of 1.2–1.5, a 10% improvement in TFLOPS/watt translates directly to lower electricity costs and cooling requirements. GB/s/watt measures memory bandwidth efficiency, which is the binding constraint for memory-bound LLM inference workloads.

When to choose each GPU

Choose NVIDIA RTX 4080 for:

  • Budget fine-tuning
  • Small model inference
  • Consumer AI

Choose NVIDIA RTX 3090 for:

  • Budget fine-tuning
  • Small model inference
  • Hobbyist AI