NVIDIA GeForce GTX 1070 Ti vs NVIDIA GeForce GTX 770 Comparison

The NVIDIA GeForce GTX 770 and the NVIDIA GeForce GTX 1070 Ti represent two distinct eras of NVIDIA's GPU architecture, separated by roughly four and a half years of process and design evolution. The data shows a stark generational divide: the GTX 1070 Ti wins all three head-to-head benchmark comparisons, with deltas ranging from 57.8% to 68.6%. While the older Kepler-based GTX 770 holds a slight edge in its aggregate benchmark percentile (57th vs. 56th), this is a reflection of the diverse test suite composition rather than raw performance parity. The GTX 1070 Ti delivers over twice the compute throughput, significantly higher memory bandwidth, and a more modern feature set, making it the clear choice for contemporary workloads. The GTX 770, however, retains relevance only in legacy DirectX 11 scenarios or for users with strict power and slot constraints, though even those advantages are marginal against the Pascal card's efficiency.

Where Each One Wins

The benchmark results are unambiguous: the GTX 1070 Ti wins every single head-to-head comparison in the data set. It secures victories in Geekbench Metal, OpenCL, and Vulkan, which are the only tests where both cards have directly comparable scores. The GTX 770 records zero wins in this comparison, meaning there is no workload category in the provided benchmarks where it outperforms its successor.

The GTX 1070 Ti's dominance is most pronounced in compute-oriented APIs. Its largest margin of victory comes in Geekbench Vulkan, where it scores 61,006 against the GTX 770's 19,166, a 68.6% difference. This suggests a significant advantage in modern, low-level graphics APIs that leverage asynchronous compute and parallel workload distribution. Similarly, in Geekbench OpenCL, the 1070 Ti's score of 49,219 dwarfs the 770's 15,707, a 68.1% gap, indicating superior general-purpose compute performance for tasks like physics simulation or video encoding.

The only area where the GTX 770 shows a comparative advantage is in its average benchmark score relative to its nearest rivals. With an average score of 14,747, it sits within 0.4% of the AMD Radeon RX 5500 XT and is only 2.1% behind the GTX 660 Ti. This indicates that within its own generation, the GTX 770 was a strong performer. However, this historical context does not translate to any win against the GTX 1070 Ti, which posts a lower average score of 14,277 but achieves that figure across a much broader and more demanding test suite, including PassMark DirectX 12 and 3DMark Steel Nomad DX12.

Architecture Differences

The architectural gap between these two cards is foundational. The GTX 770 is built on the Kepler architecture, fabricated on a 28 nm process at TSMC, packing 3,540 million transistors into a 294 mm² die. In contrast, the GTX 1070 Ti uses the Pascal architecture, fabricated on a 16 nm process, also at TSMC, housing 7,200 million transistors in a slightly larger 314 mm² die. The transition from 28 nm to 16 nm allows the Pascal chip to more than double the transistor count while maintaining a similar physical footprint, with transistor density jumping from 12.0M / mm² to 22.9M / mm².

This architectural shift yields substantial changes in core configuration. The GTX 770 features 1,536 shading units, 128 texture mapping units, and 32 ROPs. The GTX 1070 Ti scales this up to 2,432 shading units, 152 TMUs, and 64 ROPs. The doubling of ROPs is particularly significant, directly impacting pixel fill rate, which rises from 34.72 GPixel/s on the 770 to 107.7 GPixel/s on the 1070 Ti. Texture fill rate similarly improves from 138.9 GTexel/s to 255.8 GTexel/s.

Clock speeds also differ substantially. The GTX 770 runs at a base clock of 1046 MHz with a boost of 1085 MHz, while the GTX 1070 Ti operates at a base of 1607 MHz and boosts to 1683 MHz. This higher clock speed, combined with the greater core count, results in a massive FP32 compute advantage: 3.333 TFLOPS for the 770 versus 8.186 TFLOPS for the 1070 Ti. Additionally, the 1070 Ti is the only card to list an FP16 figure, at 127.9 GFLOPS (1:64), indicating a token support for half-precision that is absent on the Kepler card.

Memory architecture shows a similar generational leap. Both cards use GDDR5 on a 256-bit bus, but the 1070 Ti runs its memory at 2002 MHz (8 Gbps effective) versus 1753 MHz (7 Gbps effective) on the 770. This yields a bandwidth advantage of 256.3 GB/s versus 224.4 GB/s. More critically, the 1070 Ti offers 8 GB of VRAM compared to the 770's 2 GB, a fourfold increase that is crucial for modern high-resolution textures.

Head-to-Head Benchmarks

The direct comparisons in the data are limited to three Geekbench tests, but each tells a consistent story of generational superiority. The smallest delta is in Geekbench Metal, where the GTX 1070 Ti scores 22,180 versus the GTX 770's 9,369, a -57.8% difference from the 770's perspective. This still represents a more than twofold improvement, but the Metal API on macOS or iOS may not fully utilize the Pascal architecture's compute strengths compared to other APIs.

The gap widens considerably in Geekbench OpenCL. Here, the GTX 1070 Ti's score of 49,219 is more than three times the GTX 770's 15,707, resulting in a -68.1% delta. OpenCL is a more general-purpose compute API, and the 1070 Ti's superior FP32 throughput (8.186 TFLOPS vs. 3.333 TFLOPS) and higher memory bandwidth directly translate to this massive performance lead. This benchmark highlights the practical difference in compute capability for tasks like rendering, scientific simulations, or data processing.

The largest margin is in Geekbench Vulkan, where the GTX 1070 Ti achieves 61,006 against the GTX 770's 19,166, a -68.6% delta. Vulkan is a low-overhead API designed for modern GPUs, and the Pascal architecture's support for DirectX 12 (12_1) and Vulkan 1.4, compared to the Kepler card's DirectX 12 (11_0) and Vulkan 1.2.175, suggests better optimization for these newer workloads. The 1070 Ti's ability to nearly triple the score in this API indicates superior driver efficiency and hardware scheduling capabilities for parallel command buffers.

The GTX 1070 Ti also has additional benchmark data that the GTX 770 lacks entirely. In PassMark DirectX 11, it scores 107, while in DirectX 12 it scores 51, and in DirectX 9 it reaches 207. It also posts a PassMark G3D score of 14,674 and a GPU compute score of 7,165. These figures, absent for the GTX 770, further underscore the 1070 Ti's broader compatibility and testing coverage. In the 3DMark Steel Nomad DX12 test, the 1070 Ti scores 1,471, providing a modern gaming performance data point that the older card cannot offer.

FAQ

Q: Which card is faster in compute-heavy workloads?

A: The GTX 1070 Ti is substantially faster. It wins the Geekbench OpenCL test with a score of 49,219 versus the GTX 770's 15,707, a 68.1% difference. Its FP32 compute rating of 8.186 TFLOPS more than doubles the 770's 3.333 TFLOPS.

Q: Does the GTX 770 have any benchmark win over the GTX 1070 Ti?

A: No. In the head-to-head data provided, the GTX 1070 Ti wins all three comparisons (Geekbench Metal, OpenCL, and Vulkan). The GTX 770 records zero victories across these tests.

Q: How does the memory capacity affect real-world usage?

A: The GTX 1070 Ti offers 8 GB of VRAM, which is four times the 2 GB available on the GTX 770. This larger capacity is critical for modern games with high-resolution textures and for professional applications that require large datasets to reside in GPU memory.

Q: What is the power requirement difference between the two cards?

A: The GTX 1070 Ti has a lower thermal design power (TDP) of 180 W compared to the GTX 770's 230 W. Consequently, the suggested power supply rating is 450 W for the 1070 Ti versus 550 W for the 770. The 1070 Ti also requires only a single 8-pin power connector, while the 770 needs a 6-pin and an 8-pin connector.

Q: Which card has better support for modern graphics APIs?

A: The GTX 1070 Ti has better support. It supports DirectX 12 (12_1) and Vulkan 1.4, while the GTX 770 only supports DirectX 12 (11_0) and Vulkan 1.2.175. This gives the 1070 Ti access to more advanced rendering features in current games.

Q: How do their average benchmark scores compare?

A: The GTX 770 has a higher average benchmark score of 14,747 compared to the GTX 1070 Ti's 14,277. However, this is misleading because the 1070 Ti's average includes a much wider set of tests, including modern DirectX 12 and compute benchmarks, whereas the 770's average is based on a narrower set of Geekbench tests.

Specification Differences

| Specification | NVIDIA GeForce GTX 770 | NVIDIA GeForce GTX 1070 Ti |

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

| Architecture | Kepler | Pascal |

| Process Node | 28 nm | 16 nm |

| Transistors | 3,540 million | 7,200 million |

| Die Size | 294 mm² | 314 mm² |

| Transistor Density | 12.0M / mm² | 22.9M / mm² |

| Base Clock | 1046 MHz | 1607 MHz |

| Boost Clock | 1085 MHz | 1683 MHz |

| Memory Clock | 1753 MHz (7 Gbps effective) | 2002 MHz (8 Gbps effective) |

| Memory Size | 2 GB | 8 GB |

| Memory Bandwidth | 224.4 GB/s | 256.3 GB/s |

| Shading Units | 1536 | 2432 |

| TMUs | 128 | 152 |

| ROPs | 32 | 64 |

| Pixel Rate | 34.72 GPixel/s | 107.7 GPixel/s |

| Texture Rate | 138.9 GTexel/s | 255.8 GTexel/s |

| FP32 | 3.333 TFLOPS | 8.186 TFLOPS |

| FP16 | null | 127.9 GFLOPS (1:64) |

| TDP | 230 W | 180 W |

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

| Suggested PSU | 550 W | 450 W |

| Display Outputs | 2x DVI, 1x HDMI 1.4a, 1x DisplayPort 1.2 | 1x DVI, 1x HDMI 2.0, 3x DisplayPort 1.4a |

| DirectX Support | 12 (11_0) | 12 (12_1) |

| Vulkan Support | 1.2.175 | 1.4 |

| Dimensions (width) | 38 mm (1.5 inches) | 40 mm (1.6 inches) |

| Release Date | 2013-05-29 | 2017-11-01 |

| Predecessor | GeForce 600 | GeForce 900 |

| Successor | GeForce 900 | GeForce 20 |

| Launch MSRP | 399 USD | 399 USD |