NVIDIA GeForce RTX 3060 Ti 8GB Specifications for AI Enthusiasts
The NVIDIA GeForce RTX 3060 Ti is a mainstream gaming graphics card launched on December 2, 2020, featuring NVIDIA's Ampere architecture based on the GA104-200 GPU. It offers 4,864 CUDA cores, 152 3rd-generation tensor cores, and 38 2nd-generation RT cores with 8GB of GDDR6 memory running at 14 Gbps. The card delivers 259.2 INT8 TOPS with sparsity acceleration for AI workloads, making it suitable for INT8-quantized inference tasks. Designed to outperform the previous generation's RTX 2080 SUPER at a $399 launch price, the RTX 3060 Ti provides excellent price-to-performance for both gaming at 1440p and machine learning inference workloads. With a TDP of 200W, it uses a PCI-Express 4.0 x16 interface and features 448 GB/s memory bandwidth on a 256-bit bus. The GA104 chip is manufactured on Samsung's 8nm process with approximately 17.4 billion transistors.
- Release Date: December 2, 2020
- GPU Architecture: Ampere
- Hardware-Accelerated GEMM Operations:FP16 FP32 BF16 FP8 INT8 INT4 TF32 FP64 INT1
- CUDA Compute Capability : 8.6
Specifications for NVIDIA GeForce RTX 3060 Ti
| Raw Performance | |
|---|---|
| FP32 TFLOPs: 16.2 | |
| FP16 TFLOPs: 129.6 | |
| Tensor Core Count: 152 | |
| Memory Capacity (GB): 8 | |
| Memory Bandwidth (GB/s): 448 | |
| Int8 TOPs: 259.2 |
Real-time NVIDIA GeForce RTX 3060 Ti GPU Prices
Compare Price/Performance to other GPUs
References
- https://www.nvidia.com/content/PDF/nvidia-ampere-ga-102-gpu-architecture-whitepaper-v2.1.pdf
- https://www.nvidia.com/en-us/geforce/graphics-cards/30-series/rtx-3060-ti/
- https://www.nvidia.com/en-us/geforce/news/geforce-rtx-3060-ti-out-december-2/
- https://www.techpowerup.com/gpu-specs/geforce-rtx-3060-ti.c3681
- https://www.bestgpusforai.com/gpu-comparison/3060-ti-vs-4060