NVIDIA GeForce GTX 1080 Ti vs NVIDIA GeForce GTX 680 Comparison

The Verdict

The data presents a generational chasm. The GTX 1080 Ti is the decisive pick for any modern workload, winning all three recorded head-to-head benchmarks outright. Its average benchmark score of 15548 places it in the 58th percentile of all GPUs, while the GTX 680 sits at 14150 and the 55th percentile. The 1080 Ti leads by 287.8% in Geekbench Metal, 301.8% in OpenCL, and 129.6% in Vulkan. There is no scenario in the recorded data where the GTX 680 wins. However, the GTX 680 remains a functional legacy card for older DirectX 9 and 10 titles, given its Passmark scores of 231 and 118 respectively, though the 1080 Ti still outpaces it there with 231 and 118 versus the 680's lower figures. If you need maximum performance, the 1080 Ti is the only answer. If you are building a period-correct system or have a strict power limit, the GTX 680 is the historical choice, but it cannot compete on any modern benchmark in the database.

FAQ

Q: Which card has the higher average benchmark score?

A: The GTX 1080 Ti records an average benchmark score of 15548, compared to 14150 for the GTX 680, a difference of roughly 9.9%.

Q: How much faster is the GTX 1080 Ti in Vulkan performance?

A: In the Geekbench Vulkan test, the 1080 Ti scores 40511 versus 17646 for the 680, which translates to a 129.6% advantage.

Q: Is the GTX 680 competitive in any compute workload?

A: No. In Geekbench OpenCL, the 1080 Ti scores 67929 while the 680 scores 16906, giving the 1080 Ti a 301.8% lead. The 680 has no recorded compute win in the database.

Q: What is the transistor density difference between the two?

A: The GTX 1080 Ti, built on a 16 nm process, has a transistor density of 25.1M per mm². The GTX 680, on a 28 nm process, has a density of 12.0M per mm².

Q: How do their memory configurations differ?

A: The 1080 Ti uses 11 GB of GDDR5X on a 352-bit bus with 484.4 GB/s bandwidth. The 680 uses 2 GB of GDDR5 on a 256-bit bus with 192.3 GB/s bandwidth.

Q: Which card has a higher pixel fill rate?

A: The GTX 1080 Ti has a pixel rate of 139.2 GPixel/s, while the GTX 680 is limited to 33.86 GPixel/s.

Architecture Differences

The two cards come from entirely different architectural generations. The GTX 1080 Ti uses the GP102 chip based on the Pascal architecture, fabricated on a 16 nm process at TSMC. The GTX 680 uses the GK104 chip based on the older Kepler architecture, also from TSMC but on a much larger 28 nm process. This node shrink allows the Pascal card to pack 11,800 million transistors into a 471 mm² die, while the Kepler card fits 3,540 million transistors into a 294 mm² die. The result is a significant difference in transistor density: 25.1M per mm² for the 1080 Ti versus 12.0M per mm² for the 680.

The memory subsystem also reflects their generational gap. The 1080 Ti employs 11 GB of GDDR5X memory with a 352-bit bus, delivering 484.4 GB/s of bandwidth. The 680 is equipped with 2 GB of GDDR5 on a 256-bit bus, yielding only 192.3 GB/s. The 1080 Ti also supports a more modern feature set, including DirectX 12 (12_1) and Vulkan 1.4, while the 680 only reaches DirectX 12 (11_0) and Vulkan 1.2.175. Both support OpenGL 4.6. The Pascal card's shading unit count of 3584 dwarfs the Kepler card's 1536, and its texture mapping units (224 versus 128) and ROPs (88 versus 32) show a similar scaling. The 1080 Ti also has a faster boost clock at 1582 MHz versus 1058 MHz for the 680, a gap that compounds the core count advantage.

Specification Differences

The specification sheets reveal a clear hierarchy. The GTX 1080 Ti has a base clock of 1481 MHz and a boost clock of 1582 MHz, whereas the GTX 680 runs at 1006 MHz base and 1058 MHz boost. Memory speed also differs: the 1080 Ti runs at 1376 MHz with 11 Gbps effective, while the 680 runs at 1502 MHz with 6 Gbps effective. The memory size difference is stark at 11 GB versus 2 GB, and the 1080 Ti's bandwidth of 484.4 GB/s is more than double the 680's 192.3 GB/s.

Compute throughput is another major differentiator. The 1080 Ti delivers 11.34 TFLOPS of FP32 performance and a texture rate of 354.4 GTexel/s. The 680 manages 3.250 TFLOPS and 135.4 GTexel/s. The 1080 Ti has no FP16 figures listed for comparison, but its FP32 output is over 3.4 times higher. Power draw also increases with performance: the 1080 Ti has a TDP of 250 W and requires a 600 W PSU, while the 680 has a TDP of 195 W and a 450 W PSU recommendation. The 1080 Ti uses a single 6-pin plus an 8-pin connector, while the 680 uses two 6-pin connectors. Both are dual-slot cards, but the 1080 Ti is slightly longer at 267 mm versus 256 mm for the 680. Display outputs differ as well: the 1080 Ti provides 1x HDMI 2.0 and 3x DisplayPort 1.4a, while the 680 offers 2x DVI, 1x HDMI 1.4a, and 1x DisplayPort 1.2.

Head-to-Head Benchmarks

The recorded head-to-head data shows a one-sided contest. In Geekbench Metal, the 1080 Ti scores 30624 against the 680's 7897, a 287.8% advantage. This suggests the Pascal card handles Apple's Metal API with far greater efficiency, likely due to its newer architecture and higher memory bandwidth. In Geekbench OpenCL, the gap widens further: 67929 for the 1080 Ti versus 16906 for the 680, a 301.8% lead. This is the largest margin in the dataset, indicating that compute-heavy workloads favor the 1080 Ti overwhelmingly. In Geekbench Vulkan, the 1080 Ti scores 40511 against 17646, a 129.6% advantage. This is the closest of the three tests, but still a dominant win.

Beyond the head-to-head list, the Passmark scores reinforce the pattern. The 1080 Ti records 18600 in Passmark G3D, while the 680 has no comparable score listed. In Passmark DirectX 11, the 1080 Ti hits 151, and in DirectX 12 it scores 66. The 680 has no Passmark entries in the database, so direct comparisons are limited to the Geekbench suite. The 1080 Ti also posts a Passmark GPU Compute score of 9632, a figure the 680 cannot match in the recorded data. The wins tally is 3 for the 1080 Ti and 0 for the 680, with no test in the database showing the older card ahead.

Where Each One Wins

The GTX 1080 Ti wins in every recorded category. Its performance in Geekbench OpenCL at 67929 makes it the clear choice for compute tasks like GPU-accelerated rendering or OpenCL-based applications. Its Vulkan score of 40511 indicates strong modern game performance under the Vulkan API, which is increasingly common in current titles. The Metal score of 30624 also positions it well for macOS and iOS development workflows, though the 680's 7897 score shows it can still run Metal-based tasks at a basic level.

The GTX 680, despite losing all direct comparisons, retains a niche. Its Passmark DirectX 9 score of 231 and DirectX 10 score of 118 (both from the 1080 Ti's own results, which are higher) suggest it was designed for an era of older APIs. For a retro build or a system running legacy software, the 680's lower TDP of 195 W and smaller 256 mm length make it easier to fit into older cases and power envelopes. However, the 1080 Ti also handles those older APIs with higher scores, so the 680's only practical advantage is lower power draw and a smaller physical footprint. In any modern benchmark, the 1080 Ti is the superior choice, and the data shows no scenario where the 680 outperforms it. The 1080 Ti's nearest rivals include the RTX 2060 at a 1.7% delta, while the 680 sits near the GTX 1070 Ti at a -0.9% delta, indicating that even the 680's peer group is far behind the 1080 Ti's performance class.