NVIDIA GeForce GTX 1060 6 GB vs NVIDIA GeForce RTX 2070 Comparison

FAQ

Q: How do the two cards compare in raw compute performance?

A: The RTX 2070 delivers 7.465 TFLOPS FP32 versus 4.375 TFLOPS for the GTX 1060 6 GB. That is a 70.6% higher FP32 throughput, and in FP16 the gap is far larger: 14.93 TFLOPS versus 68.36 GFLOPS, a 1:64 ratio versus a 2:1 ratio.

Q: Which card has more memory, and does it matter?

A: The RTX 2070 has 8 GB GDDR6 on a 256-bit bus, yielding 448.0 GB/s bandwidth. The GTX 1060 6 GB has 6 GB GDDR5 on a 192-bit bus, yielding 192.2 GB/s. The RTX 2070 offers 2.33 times the memory bandwidth.

Q: What are the architecture generations?

A: The GTX 1060 6 GB is Pascal (GP106) on TSMC 16 nm, released in July 2016. The RTX 2070 is Turing (TU106) on TSMC 12 nm, released in October 2018. The RTX 2070 also adds dedicated ray tracing cores (36) and tensor cores (288), neither of which exist on the GTX 1060.

Q: Which card performs better in the recorded benchmark suite?

A: The GTX 1060 6 GB has a higher average benchmark score of 21856 across its tests, while the RTX 2070 averages 18789. However, in the direct head-to-head tests, the RTX 2070 wins both: OpenCL (79966 vs 32322) and Vulkan (82521 vs 18659).

Q: How do their power requirements differ?

A: The GTX 1060 6 GB has a TDP of 120 W with a 1x 6-pin connector and a suggested PSU of 300 W. The RTX 2070 has a TDP of 175 W, requires a 1x 8-pin connector, and suggests a 450 W PSU.

Q: Do they support the same DirectX feature level?

A: No. The GTX 1060 6 GB supports DirectX 12 (12_1), while the RTX 2070 supports DirectX 12 Ultimate (12_2). Both support OpenGL 4.6 and Vulkan 1.4.

Architecture Differences

The two cards represent distinct NVIDIA architectures. The GTX 1060 6 GB is built on the Pascal architecture, specifically the GP106 chip, fabricated by TSMC on a 16 nm process. The RTX 2070 uses the Turing architecture with the TU106 chip, also from TSMC but on a 12 nm node. The process shrink allows the RTX 2070 to pack significantly more hardware: 10,800 million transistors on a 445 mm² die, versus 4,400 million on 200 mm² for the GTX 1060. Transistor density rises from 22.0M per mm² to 24.3M per mm².

The most significant architectural addition on the RTX 2070 is the inclusion of 36 ray tracing cores and 288 tensor cores. The GTX 1060 6 GB has none. These dedicated units enable hardware-accelerated ray tracing and AI-based tensor operations, which are entirely absent from the Pascal card. The RTX 2070 also doubles the FP16 throughput ratio: it processes FP16 at 14.93 TFLOPS (2:1 ratio), whereas the GTX 1060 6 GB manages only 68.36 GFLOPS (1:64 ratio), meaning FP16 is a heavy penalty on the older card.

The memory subsystem differs fundamentally. The GTX 1060 6 GB uses GDDR5 at 2002 MHz (8 Gbps effective) across a 192-bit bus. The RTX 2070 uses GDDR6 at 1750 MHz (14 Gbps effective) across a 256-bit bus. The result is a bandwidth jump from 192.2 GB/s to 448.0 GB/s. The RTX 2070 also doubles the texture mapping units (144 vs 80) and increases render output units from 48 to 64.

The process node and transistor count differences have a direct impact on power efficiency and thermals. The RTX 2070 carries a TDP of 175 W versus 120 W for the GTX 1060 6 GB, despite the newer node. The larger die and higher transistor count explain this. The RTX 2070 requires a single 8-pin power connector and a 450 W suggested PSU, while the GTX 1060 6 GB uses a single 6-pin and a 300 W PSU.

DirectX support also diverges: the GTX 1060 6 GB is limited to DirectX 12 (12_1), while the RTX 2070 reaches DirectX 12 Ultimate (12_2). Both cards support OpenGL 4.6 and Vulkan 1.4.

The Verdict

The data presents a clear split. For raw compute and modern feature support, the RTX 2070 is the stronger card. Its FP32 throughput is 70.6% higher, its memory bandwidth is 2.33 times higher, and it introduces ray tracing and tensor cores. In the two direct head-to-head benchmarks, the RTX 2070 wins decisively: 79966 versus 32322 in OpenCL (a 59.6% advantage) and 82521 versus 18659 in Vulkan (a 77.4% advantage). Any workload that leverages these capabilities will favor the RTX 2070.

However, the GTX 1060 6 GB holds a higher average benchmark score overall (21856 vs 18789) and a higher percentile rank among all GPUs (67th vs 63rd). Its nearest rivals include the AMD Radeon RX 5700 (22170, 1.4% higher) and the AMD Radeon RX 6700S (22154, 1.3% higher), showing it remains competitive in its peer group. The RTX 2070 sits near the NVIDIA Tesla K80 (18866, 0.4% higher) and the AMD Radeon Pro 5700 XT (18685, 0.6% lower), indicating it is more of a mid-pack performer in the broader database.

Who should pick which? Users prioritizing compatibility with older DirectX 12 (12_1) titles, lower power draw (120 W vs 175 W), and a smaller PSU footprint (300 W vs 450 W) may find the GTX 1060 6 GB sufficient. Its 6 GB of GDDR5 memory at 192.2 GB/s is adequate for less demanding workloads. On the other hand, anyone targeting DirectX 12 Ultimate features, ray tracing, or compute-heavy tasks that benefit from FP16 at a 2:1 ratio should choose the RTX 2070. The 8 GB GDDR6 frame buffer and 448.0 GB/s bandwidth provide a substantial headroom for higher resolutions and texture-heavy scenes.

The production status for both is end-of-life, but the RTX 2070 benefits from a newer generation (GeForce 20) and a later release date (October 2018 vs July 2016). The GTX 1060 6 GB belongs to GeForce 10, with its successor listed as GeForce 20, which is precisely the RTX 2070's generation. In essence, the RTX 2070 is the direct successor class, and the benchmark delta supports that positioning.

Specification Differences

The table below lists only the fields where the two cards differ, based on the recorded data.

| Field | NVIDIA GeForce GTX 1060 6 GB | NVIDIA GeForce RTX 2070 |

|-------|-----------------------------|-------------------------|

| Series | GeForce 10-series | GeForce 20-series |

| Chip | GP106 | TU106 |

| Architecture | Pascal | Turing |

| Generation | GeForce 10 | GeForce 20 |

| Process Node | 16 nm | 12 nm |

| Transistors | 4,400 million | 10,800 million |

| Die Size | 200 mm² | 445 mm² |

| Transistor Density | 22.0M / mm² | 24.3M / mm² |

| Base Clock | 1506 MHz | 1410 MHz |

| Boost Clock | 1709 MHz | 1620 MHz |

| Memory Clock | 2002 MHz, 8 Gbps effective | 1750 MHz, 14 Gbps effective |

| Memory Size | 6 GB | 8 GB |

| Memory Type | GDDR5 | GDDR6 |

| Memory Bus Width | 192 bit | 256 bit |

| Memory Bandwidth | 192.2 GB/s | 448.0 GB/s |

| Shading Units | 1280 | 2304 |

| TMUs | 80 | 144 |

| ROPs | 48 | 64 |

| RT Cores | None | 36 |

| Tensor Cores | None | 288 |

| Pixel Rate | 82.03 GPixel/s | 103.7 GPixel/s |

| Texture Rate | 136.7 GTexel/s | 233.3 GTexel/s |

| FP32 | 4.375 TFLOPS | 7.465 TFLOPS |

| FP16 | 68.36 GFLOPS (1:64) | 14.93 TFLOPS (2:1) |

| TDP | 120 W | 175 W |

| Power Connectors | 1x 6-pin | 1x 8-pin |

| Suggested PSU | 300 W | 450 W |

| Display Outputs | 1x DVI, 1x HDMI 2.0, 3x DisplayPort 1.4a | 1x DVI, 1x HDMI 2.0, 2x DisplayPort 1.4a, 1x USB Type-C |

| DirectX | 12 (12_1) | 12 Ultimate (12_2) |

| Length | 250 mm (9.8 inches) | 229 mm (9 inches) |

| Height | 111 mm (4.4 inches) | 113 mm (4.4 inches) |

| Width | Not specified | 35 mm (1.4 inches) |

| Release Date | 2016-07-18 | 2018-10-16 |

| Predecessor | GeForce 900 | GeForce 10 |

| Successor | GeForce 20 | GeForce 30 |

| Launch MSRP | 299 USD | 499 USD |

Head-to-Head Benchmarks

The database records two direct head-to-head tests between the GTX 1060 6 GB and the RTX 2070. Both are won by the RTX 2070, with no wins for the GTX 1060 6 GB in this subset.

In Geekbench OpenCL, the RTX 2070 scores 79966 against 32322 for the GTX 1060 6 GB. The delta percentage is -59.6% from the perspective of the GTX 1060, meaning the RTX 2070 outperforms it by 147.4% (since 79966 / 32322 = 2.47). This is a massive generational leap in compute throughput, consistent with the FP32 and FP16 figures: the RTX 2070 has 1.71 times the FP32 throughput and over 200 times the FP16 throughput.

In Geekbench Vulkan, the RTX 2070 scores 82521 versus 18659 for the GTX 1060 6 GB. The delta is -77.4%, which translates to a 4.42 times advantage for the RTX 2070 (82521 / 18659). This larger margin suggests that Vulkan workloads, which may leverage async compute or newer driver optimizations, benefit disproportionately from the Turing architecture's hardware features.

Looking at the broader benchmark suite, the GTX 1060 6 GB has no recorded wins in the head-to-head section, but its average benchmark score of 21856 actually exceeds the RTX 2070's 18789. This discrepancy arises because the GTX 1060's benchmark set (Geekbench Metal, OpenCL, Vulkan) may be more favorable to its architecture, while the RTX 2070's set includes Passmark DirectX 9/10/11/12, G2D, G3D, and GPU compute tests, plus a 3DMark Steel Nomad DX12 test. The RTX 2070 scores 16094 in Passmark G3D, 6411 in GPU compute, and 1569 in 3DMark Steel Nomad, but these do not directly compare to the GTX 1060's tests.

The nearest rivals further contextualize each card. The GTX 1060 6 GB sits within 2.2% of the NVIDIA RTX A4000 Mobile (21379) and 2.1% of the NVIDIA Quadro RTX 5000 (21629), with both opponents slightly lower. The RTX 2070 sits within 0.9% of the AMD Radeon RX 560X (18626) and 0.6% of the AMD Radeon Pro 5700 XT (18685), indicating a tighter cluster around its average.

In practical terms, the biggest win for the RTX 2070 is in Vulkan, where it more than quadruples the GTX 1060's score. The smallest, yet still decisive, win is in OpenCL, where it is roughly 2.5 times faster. For any user moving from a GTX 1060 6 GB to an RTX 2070, the recorded data indicates a dramatic uplift in compute-oriented tasks, at the cost of higher power draw and a larger PSU requirement. The GTX 1060 6 GB, meanwhile, offers a lower power envelope and a higher average score in its own test set, but it cannot match the RTX 2070 in any direct comparison recorded here.