| 规格 | NVIDIA GeForce RTX 3060 12GB | NVIDIA A30 |
|---|
| 架构 | Ampere (GA106-300, 28 SM, 3MB L2) | NVIDIA Ampere |
| 显存 | 12GB GDDR6 | 24 GB HBM2 |
| 显存带宽 | 360GB/s (15Gbps x 192bit) | 933 GB/s |
| 显存位宽 | 192bit | — |
| 算力 | FP32 12.74 TFLOPS (Wikipedia按Boost 1777MHz核算, 常引作12.7; Base 9.46) / Tensor FP16 51.2 TFLOPS [102.4稀疏] | — |
| CUDA核心 | 3584 (28 SM) / Tensor Core 112 (三代) / RT Core 28 (二代) | — |
| 功耗 | 170W TGP (建议系统电源550W, 1x PCIe 8-pin附1x 12-pin转接, 最高93°C) | — |
| 互联 | PCIe Gen4 x16, 无NVLink (官方页NVLink(SLI-Ready)标注'-') | PCIe Gen4: 64GB/s / Third-gen NVLINK: 200GB/s |
| 发布年 | 2021 (CES 2021发布, 2021-02-25开售, MSRP $329) | — |
| 频率 | Base 1.32GHz / Boost 1.78GHz | — |
| 工艺 | Samsung 8N, 132.5亿晶体管, 276mm² (GA106满血30 SM:3840CUDA, 3060启用28) | — |
| 媒体引擎 | 第7代NVENC / 第5代NVDEC (官方页标注NVDEC为Decoder) | — |
| 显示输出 | HDMI 2.1 + DP 1.4a x3, 4屏, HDCP 2.3, 7680x4320 | — |
| 形态 | FE双槽 242x112mm, 供电 1x8pin(附1x 12-pin转接), 需550W电源 | Dual-slot, full-height, full-length (FHFL) |
| CUDA算力等级 | Compute Capability 8.6 | — |
| FP32算力 | — | 10.3 teraFLOPS |
| FP64算力 | — | 5.2 teraFLOPS |
| TF32算力 | — | 82 teraFLOPS | 165 teraFLOPS |
| FP16算力 | — | 165 teraFLOPS | 330 teraFLOPS |
| INT8算力 | — | 330 TOPS | 661 TOPS |
| 功耗TDP | — | 165 W |
| MIG | — | 4 GPU instances @ 6GB each |