Flopper.io GPU Comparison Sheet
September 4, 2026
flopper.io/compare/nvidia-rtx-4000-ada-20gb-vs-nvidia-l40-48gbNVIDIA RTX 4000 Ada 20GB vs NVIDIA L40 48GB
Side by side datacenter GPU specifications from the Flopper.io database.
Overview
| NVIDIA RTX 4000 Ada 20GB | NVIDIA L40 48GB | |
|---|---|---|
| FP32 TFLOPS | 26.7 | 90.5 |
| VRAM | 20 GB | 48 GB |
| Memory Bandwidth | 360 GB/s | 864 GB/s |
| TDP | 130 W | 300 W |
Performance by Precision
| Precision | NVIDIA RTX 4000 Ada 20GB peak TFLOPS (dense) | NVIDIA L40 48GB peak TFLOPS (dense) |
|---|---|---|
| FP64 | 0.4 | — |
| FP32 | 26.7 | 90.5 |
| TF32 | — | 90.5 |
| BF16 | — | 181.1 |
| FP16 | — | 181.1 |
| FP8 | 213.8 | 362.0 |
| FP6 | — | — |
| FP4 | — | — |
| INT8 | — | 362.0 |
| INT4 | — | 724.0 |
Specifications
| NVIDIA RTX 4000 Ada 20GB | NVIDIA L40 48GB | |
|---|---|---|
| Architecture | Ada Lovelace | Ada Lovelace |
| Process Node | 4nm | 4nm |
| Launch Year | 2023 | 2023 |
| Form Factor | PCIe | PCIe |
| Memory Type | GDDR6 | GDDR6 |
| Interface Width | 160-bit | — |
| Interconnect | — | — |
| TDP | 130 W | 300 W |
| FP32 Efficiency | 0.21 TFLOPS/W | 0.30 TFLOPS/W |