NVIDIA RTX 4000 SFF Ada 20GB vs NVIDIA L40S 48GB

The L40S delivers 4.8x the FP8 throughput of the RTX 4000 SFF Ada (733 vs 153 TFLOPS dense). The L40S also carries 28 GB more memory (48 GB vs 20 GB).

RTX 4000 SFF Ada: Ada Lovelace, 2023 L40S: Ada Lovelace, 2023
FP8 dense lead
4.8x
L40S: 733 vs 153 TFLOPS
VRAM
20 vs 48 GB
28 GB more for the L40S
Bandwidth
280 GB/s vs 864 GB/s
L40S moves data faster
TDP
70 W vs 350 W
RTX 4000 SFF Ada draws 280 W less

Specifications

RTX 4000 SFF Ada
L40S
Architecture
Ada LovelaceAda Lovelace
Launch Year
20232023
Form Factor
PCIePCIe
VRAM
20 GB48 GB
Memory Bandwidth
280 GB/s864 GB/s
TDP
70 W350 W
Process Node
4nm4nm

Performance (TFLOPS)

RTX 4000 SFF Ada
L40S
FP64
0.3 TFlops 1.4 TFlops
FP32
19.2 TFlops 91.6 TFlops
TF32
No verified data available183.0 TFlops
366.0 TFLOPS sparse
BF16
No verified data available362.1 TFlops
733.0 TFLOPS sparse
FP16
No verified data available362.1 TFlops
733.0 TFLOPS sparse
FP8
153.4 TFlops
306.8 TFLOPS sparse
733.0 TFlops
1466.0 TFLOPS sparse
FP6
No verified data available No verified data available
FP4
No verified data available No verified data available
INT8
No verified data available733.0 TFlops
1466.0 TOPS sparse

FLOPS by Precision

What actually differs

Both GPUs launched in 2023: the RTX 4000 SFF Ada on Ada Lovelace and the L40S on Ada Lovelace.

Dense throughput for the RTX 4000 SFF Ada against the L40S: FP64 0.3 vs 1.4 TFLOPS, FP32 19 vs 92 TFLOPS, FP8 153 vs 733 TFLOPS. Sparse figures, where the vendor publishes them, appear under each dense number in the performance table.

Memory is 20 GB against 48 GB, fed at 280 GB/s versus 864 GB/s. More memory per GPU means larger models fit before you have to shard across cards, which often matters more than raw TFLOPS for inference.

Power budgets are 70 W for the RTX 4000 SFF Ada and 350 W for the L40S. At FP32 that works out to 0.27 against 0.26 TFLOPS per watt.

For LLM training and inference, weight the FP8 and FP16 rows and memory capacity most heavily. For scientific and HPC workloads, the FP64 row is the one to read.

Frequently asked questions

Is the RTX 4000 SFF Ada faster than the L40S?

At FP8 precision the L40S reaches 733 TFLOPS dense against 153 TFLOPS for the RTX 4000 SFF Ada. The performance table on this page lists every published precision for both GPUs.

Which has more memory, the RTX 4000 SFF Ada or the L40S?

The L40S carries 48 GB of VRAM versus 20 GB for the RTX 4000 SFF Ada. Memory bandwidth is 280 GB/s for the RTX 4000 SFF Ada and 864 GB/s for the L40S.

How much power do the RTX 4000 SFF Ada and the L40S draw?

The RTX 4000 SFF Ada is rated at 70 W TDP and the L40S at 350 W. On FP32 throughput per watt, the RTX 4000 SFF Ada is the more efficient part.

Can I rent the RTX 4000 SFF Ada or the L40S in the cloud?

Yes. Live cloud listings tracked by Flopper start at $0.18 per GPU hour. The rental pricing section on this page lists current providers and rates for both GPUs.

Where to Rent

NVIDIA RTX 4000 SFF Ada 20GB

ProviderConfigurationPrice/GPU-hrChecked
RunPod
1× RTX 4000 SFF Ada 20GB Community $0.18 1h ago View →
RunPod
1× RTX 4000 SFF Ada 20GB $0.44 1h ago View →
All RTX 4000 SFF Ada listings and price history →

NVIDIA L40S 48GB

ProviderConfigurationPrice/GPU-hrChecked
Vast.ai
1× L40S 48GB Community $0.60 1h ago View →
DataCrunch
1× L40S 48GB Spot $0.69 1h ago View →
Nebius
1× L40S 48GB Spot $0.74 1h ago View →
RunPod
1× L40S 48GB Community $0.79 1h ago View →
Cudo Compute
1× L40S 48GB $0.87 1h ago View →
Massed Compute
1× L40S 48GB $0.88 1h ago View →
CoreWeave
8× L40S 48GB Spot $0.98 1h ago View →
RunPod
1× L40S 48GB $1.09 1h ago View →
DataCrunch
1× L40S 48GB $1.37 1h ago View →
AceCloud
1× L40S 48GB $1.50 1h ago View →
Crusoe
1× L40S 48GB $1.50 1h ago View →
Nebius
1× L40S 48GB $1.55 1h ago View →
DigitalOcean
1× L40S 48GB $1.57 1h ago View →
Vultr
1× L40S 48GB $1.67 1h ago View →
OVHcloud
4× L40S 48GB $1.80 1h ago View →
CoreWeave
8× L40S 48GB $2.25 1h ago View →
All L40S listings and price history →

Get Comparison Updates

New GPUs added weekly. Be the first to see how they compare.