100% Original Wholesale Brand new ada lovelace type TU106-400 TU106/200 TU106-410
Here’s a detailed comparison table for the TU106-400, TU106-200, and TU106-410 GPU variants, following NVIDIA’s Turing-era segmentation for mid-range GPUs like the RTX 2060 series:
TU106 GPU Variants Comparison Table
(RTX 2060 Super / RTX 2060 / OEM Models)
| Specification | TU106-400Â (e.g., RTX 2060 Super) | TU106-200Â (e.g., RTX 2060) | TU106-410Â (e.g., OEM/Refresh) | Difference |
|---|---|---|---|---|
| Die Variant | Full TU106 (400-A1) | Cut-down TU106 (200-A1) | Refreshed/OEM variant (410-A1) | |
| CUDA Cores | 2,176 | 1,920 | 2,176 (same as -400) | +13% vs. 200 |
| SMs (Streaming Multiprocessors) | 34 | 30 | 34 | +4 SMs vs. 200 |
| Base Clock | 1,470 MHz | 1,365 MHz | ~1,410 MHz (OEM-adjusted) | +3% vs. 200 |
| Boost Clock | 1,650 MHz | 1,680 MHz | ~1,710 MHz (OEM-boosted) | +2% vs. 200 |
| Memory Size | 8GB GDDR6 | 6GB GDDR6 | 8GB GDDR6 | +33% vs. 200 |
| Memory Bus Width | 256-bit | 192-bit | 256-bit | Wider bus vs. 200 |
| Memory Bandwidth | 448 GB/s | 336 GB/s | 448 GB/s | +33% vs. 200 |
| RT Cores | 34 (2nd Gen) | 30 | 34 | +13% vs. 200 |
| Tensor Cores | 272 (3rd Gen) | 240 | 272 | +13% vs. 200 |
| TDP | 175W | 160W | ~170W (OEM-adjusted) | +6% vs. 200 |
| Typical SKU | RTX 2060 Super | RTX 2060 |
TU106 GPU Variants Technical Descriptions
1. TU106-400 (Full Die – RTX 2060 Super)
The TU106-400 represents the fully enabled implementation of NVIDIA’s mid-range Turing GPU, featuring:
- 2,176 CUDA cores (34 SMs)
- 34 RT cores (2nd generation)
- 272 Tensor cores (3rd generation)
- 8GB GDDR6 memory on a 256-bit bus (448 GB/s bandwidth)
This configuration delivers excellent 1440p gaming performance with:
- Base clock:Â 1,470 MHz
- Boost clock:Â 1,650 MHz
- TDP:Â 175W
The TU106-400 powers the RTX 2060 Super, offering performance comparable to the original RTX 2070 while maintaining better power efficiency. It supports all Turing architecture features including:
- Real-time ray tracing
- DLSS (Deep Learning Super Sampling)
- NVENC improved video encoding
2. TU106-200 (Cut-Down Die – RTX 2060)
The TU106-200 is a cost-reduced variant with:
- 1,920 CUDA cores (30 SMs)
- 30 RT cores
- 240 Tensor cores
- 6GB GDDR6 memory on a 192-bit bus (336 GB/s bandwidth)
Performance characteristics:
- Base clock:Â 1,365 MHz
- Boost clock:Â 1,680 MHz
- TDP:Â 160W
As the GPU behind the original RTX 2060, the TU106-200 brought ray tracing to the mainstream market. While it has:
- 12% fewer cores than the -400 variant
- 25% less memory bandwidth
It remains capable of 1080p ray traced gaming and benefits from Turing’s architectural improvements.
3. TU106-410 (OEM Refresh Variant)
The TU106-410 serves as an OEM-focused refresh with:
- Full 2,176 CUDA cores (same as -400)
- 34 RT cores
- 272 Tensor cores
- 8GB GDDR6 memory (256-bit bus)
Key differentiators:
- Slightly adjusted clocks for OEM systems
- Typically appears in pre-built systems (Dell, HP, etc.)
- May feature small clock speed bumps over reference designs
Reviews
There are no reviews yet.