Flopper.io GPU Comparison Sheet
NVIDIA A100 PCIe 40GB vs NVIDIA GeForce RTX 3060 Ti 8GB
Side by side datacenter GPU specifications from the Flopper.io database.
Overview
| NVIDIA A100 PCIe 40GB | NVIDIA GeForce RTX 3060 Ti 8GB | |
|---|---|---|
| FP32 TFLOPS | 19.5 | 16.2 |
| Memory | 40 GB | 8 GB |
| Memory Bandwidth | 1.6 TB/s | — |
| TDP | 250 W | 200 W |
Performance by Precision
| Precision | NVIDIA A100 PCIe 40GB peak TFLOPS (dense) | NVIDIA GeForce RTX 3060 Ti 8GB peak TFLOPS (dense) |
|---|---|---|
| FP64 | 9.7 | 0.3 |
| FP32 | 19.5 | 16.2 |
| TF32 | 156.0 | 16.2 |
| BF16 | 312.0 | 32.4 |
| FP16 | 312.0 | 32.4 |
| FP8 | — | — |
| FP6 | — | — |
| FP4 | — | — |
| INT8 | 624.0 | 129.6 |
| INT4 | 1248.0 | 259.1 |
Specifications
| NVIDIA A100 PCIe 40GB | NVIDIA GeForce RTX 3060 Ti 8GB | |
|---|---|---|
| Architecture | Ampere | Ampere |
| Process Node | 7nm | 8nm |
| Launch Year | 2020 | 2020 |
| Form Factor | PCIe | PCIe |
| Memory Type | — | GDDR6 |
| Interface Width | — | 256-bit |
| Interconnect | — | — |
| TDP | 250 W | 200 W |
| FP32 Efficiency | 0.08 TFLOPS/W | 0.08 TFLOPS/W |