A100 SXM4 80GB vs L40S
Compare NVIDIA A100 SXM4 80GB and NVIDIA L40S specs, performance, and cloud pricing
A100 SXM4 80GB
80GB
From $1.10/hr
L40S
48GB
From $0.820/hr
Architecture
Ampere
vs Ada Lovelace
FP16 Gap
1.7x
A100 SXM4 80GB leads
| Specification | A100 SXM4 80GB | L40S |
|---|---|---|
| VRAM | 80 GB | 48 GB |
| VRAM Type | HBM2e | GDDR6X |
| FP16 TFLOPS | 624 TFLOPS | 366.5 TFLOPS |
| FP8 TFLOPS | N/A | 733 TFLOPS |
| Memory Bandwidth | 2.0 TB/s | 864 GB/s |
| TDP | 400W | 350W |
| Interconnect | NVLink 3 | PCIe Gen4 |
| Architecture | Ampere | Ada Lovelace |
Price Comparison
| Metric | A100 SXM4 80GB | L40S |
|---|---|---|
| Cheapest On-Demand | $1.10/hr | $0.820/hr |
| Cheapest Spot | $0.760/hr | $0.440/hr |
| Providers Available | 6 | 5 |
Verdict
Best for Training
NVIDIA A100 SXM4 80GB
624 TFLOPS FP16 with 80GB VRAM
Best Value
NVIDIA A100 SXM4 80GB
567 TFLOPS per $/hr
Best for Inference
NVIDIA L40S
733 TFLOPS FP8/FP16
Use-Case Recommendations
Large-Scale Training
Training LLMs and large multi-modal models
Winner
A100 SXM4 80GB
624 TFLOPS FP16 with 80GB HBM2e provides the best training throughput.
Inference at Scale
Deploying models in production for real-time inference
Winner
L40S
733 TFLOPS FP8/FP16 gives superior inference throughput.
Budget-Conscious Workloads
Getting the best performance per dollar
Winner
A100 SXM4 80GB
Starting at $1.10/hr delivers the best TFLOPS per dollar.